{ "success": true, "total_models": 210, "ecosystem": "AKI 210-Model Live Frontier Intelligence", "edition": "September 2026 Live Edge Release", "governance": "Zero-Incentive-Protocol (ZIP-1.0)", "last_updated": "2026-09-18T21:00:00Z", "stats": { "proprietary_models": 89, "open_weights_models": 121, "top_swe_bench_verified": "Anthropic Claude Fable 5.1 (82.6%)", "top_osworld_desktop_autonomy": "OpenAI GPT-6 Astra (72.6%)", "top_gpqa_stem": "OpenAI GPT-5.6 Sol (91.2%)", "lowest_cache_read_pricing": "Anthropic Claude Fable 5.1 ($0.25/M)", "longest_context_window": "Google DeepMind Gemini 3.0 Pro & xAI Grok 4.5 (2M tokens)" }, "models": [ { "rank": 1, "model_id": "claude-fable-5.1", "name": "Claude Fable 5.1", "developer": "Anthropic", "license_type": "proprietary", "composite_iq_score": 99.8, "reasoning_score": 99.6, "agent_code_score": 82.6, "stem_gpqa_score": 89.4, "osworld_score": 70.4, "human_elo": 1410, "multimodal_score": 98.2, "context_window": "500k", "pricing_input_per_m": 10, "pricing_output_per_m": 50, "cache_read_per_m": 0.25, "reality_gap": 6.8, "category": "Frontier Reasoning", "key_moat": "SOTA Coding & Science, Constitutional Hybrid Thought with sub-second cache reads", "release_date": "2026-09-02", "official_url": "https://claude.ai", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 2, "model_id": "gpt-6-astra", "name": "OpenAI GPT-6 Astra", "developer": "OpenAI", "license_type": "proprietary", "composite_iq_score": 99.5, "reasoning_score": 99.4, "agent_code_score": 81.4, "stem_gpqa_score": 90.1, "osworld_score": 72.6, "human_elo": 1405, "multimodal_score": 98.9, "context_window": "1M", "pricing_input_per_m": 12, "pricing_output_per_m": 48, "cache_read_per_m": 0.5, "reality_gap": 16.4, "category": "Coding & Agentic", "key_moat": "72.6% OSWorld 2.0 benchmark leader, Stargate cluster inference, Multimodal Operator", "release_date": "2026-09-03", "official_url": "https://openai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 3, "model_id": "claude-opus-4.5", "name": "Claude Opus 4.5", "developer": "Anthropic", "license_type": "proprietary", "composite_iq_score": 99.2, "reasoning_score": 99.1, "agent_code_score": 80.9, "stem_gpqa_score": 88.6, "osworld_score": 68.2, "human_elo": 1398, "multimodal_score": 97.8, "context_window": "500k", "pricing_input_per_m": 15, "pricing_output_per_m": 75, "cache_read_per_m": 0.75, "reality_gap": 7.2, "category": "Frontier Reasoning", "key_moat": "80.9% SWE-bench Verified, ultra-deep epistemic reasoning and formal theorem synthesis", "release_date": "2026-08-15", "official_url": "https://claude.ai", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 4, "model_id": "grok-4.5", "name": "Grok 4.5", "developer": "xAI", "license_type": "proprietary", "composite_iq_score": 98.9, "reasoning_score": 98.7, "agent_code_score": 77, "stem_gpqa_score": 86.8, "osworld_score": 64.5, "human_elo": 1390, "multimodal_score": 95.5, "context_window": "2M", "pricing_input_per_m": 5, "pricing_output_per_m": 15, "cache_read_per_m": 0.3, "reality_gap": 11.2, "category": "Coding & Agentic", "key_moat": "SOTA TerminalBench & SWE Marathon leader, Long-Context Agentic Engineering on Colossus 200k", "release_date": "2026-08-20", "official_url": "https://x.ai", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 5, "model_id": "gemini-3-pro", "name": "Gemini 3.0 Pro", "developer": "Google DeepMind", "license_type": "proprietary", "composite_iq_score": 98.7, "reasoning_score": 98.5, "agent_code_score": 78.5, "stem_gpqa_score": 87.2, "osworld_score": 66, "human_elo": 1388, "multimodal_score": 99.4, "context_window": "2M", "pricing_input_per_m": 2.5, "pricing_output_per_m": 10, "cache_read_per_m": 0.2, "reality_gap": 9.5, "category": "Multimodal & Vision", "key_moat": "Frontier Multimodal & Speed, 2M native context with video/audio continuous understanding", "release_date": "2026-09-01", "official_url": "https://gemini.google.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 6, "model_id": "gpt-5.6-sol", "name": "OpenAI GPT-5.6 Sol", "developer": "OpenAI", "license_type": "proprietary", "composite_iq_score": 98.5, "reasoning_score": 99.3, "agent_code_score": 79.2, "stem_gpqa_score": 91.2, "osworld_score": 63, "human_elo": 1385, "multimodal_score": 96, "context_window": "256k", "pricing_input_per_m": 10, "pricing_output_per_m": 40, "cache_read_per_m": 0.4, "reality_gap": 13.5, "category": "Enterprise & Math", "key_moat": "Frontier Math & Reasoning, Olympiad gold-tier formal verification & Lean integration", "release_date": "2026-08-25", "official_url": "https://openai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 7, "model_id": "gemini-2.5-flash", "name": "Gemini 2.5 Flash", "developer": "Google DeepMind", "license_type": "proprietary", "composite_iq_score": 97.4, "reasoning_score": 96.8, "agent_code_score": 74, "stem_gpqa_score": 82.4, "osworld_score": 59, "human_elo": 1362, "multimodal_score": 98, "context_window": "1M", "pricing_input_per_m": 0.15, "pricing_output_per_m": 0.6, "cache_read_per_m": 0.04, "reality_gap": 5.4, "category": "Efficiency & Edge", "key_moat": "Frontier Multimodal & Speed at sub-$0.20 economics with 1M native context", "release_date": "2026-08-10", "official_url": "https://gemini.google.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 8, "model_id": "deepseek-v4", "name": "DeepSeek-V4", "developer": "DeepSeek AI", "license_type": "open_weights", "composite_iq_score": 98.2, "reasoning_score": 98.4, "agent_code_score": 76.4, "stem_gpqa_score": 85, "osworld_score": 58.5, "human_elo": 1382, "multimodal_score": 93, "context_window": "256k", "pricing_input_per_m": 0.2, "pricing_output_per_m": 0.6, "cache_read_per_m": 0.05, "reality_gap": 4.8, "category": "Open-Weights SOTA", "key_moat": "Open-weights SOTA sparse MoE with multi-head latent attention (MLA) and MIT license", "release_date": "2026-08-01", "official_url": "https://deepseek.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 9, "model_id": "qwen-3.8-max", "name": "Qwen 3.8 Max", "developer": "Alibaba Cloud", "license_type": "proprietary", "composite_iq_score": 98, "reasoning_score": 98.1, "agent_code_score": 75.8, "stem_gpqa_score": 84.5, "osworld_score": 60.2, "human_elo": 1379, "multimodal_score": 96.5, "context_window": "256k", "pricing_input_per_m": 1.2, "pricing_output_per_m": 4.8, "cache_read_per_m": 0.15, "reality_gap": 5.9, "category": "Frontier Reasoning", "key_moat": "Sovereign Asian frontier foundation model with comprehensive STEM and bilingual tool calling", "release_date": "2026-08-12", "official_url": "https://alibabacloud.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 10, "model_id": "kimi-k3", "name": "Kimi K3", "developer": "Moonshot AI", "license_type": "proprietary", "composite_iq_score": 97.8, "reasoning_score": 97.9, "agent_code_score": 74.5, "stem_gpqa_score": 83.2, "osworld_score": 59, "human_elo": 1374, "multimodal_score": 94, "context_window": "500k", "pricing_input_per_m": 0.8, "pricing_output_per_m": 2.4, "cache_read_per_m": 0.1, "reality_gap": 6.2, "category": "Coding & Agentic", "key_moat": "Extended agentic reasoning with autonomous self-critique loops and deep document synthesis", "release_date": "2026-08-18", "official_url": "https://moonshot.cn", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 11, "model_id": "openai-o3", "name": "OpenAI o3", "developer": "OpenAI", "license_type": "proprietary", "composite_iq_score": 97.6, "reasoning_score": 98.3, "agent_code_score": 73.8, "stem_gpqa_score": 89.2, "osworld_score": 55, "human_elo": 1355, "multimodal_score": 93.8, "context_window": "200k", "pricing_input_per_m": 10, "pricing_output_per_m": 40, "cache_read_per_m": 1, "reality_gap": 8.4, "category": "Enterprise & Math", "key_moat": "Test-time compute scaling Olympiad reasoning", "release_date": "2026-06-15", "official_url": "https://openai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 12, "model_id": "openai-o3-mini", "name": "OpenAI o3-mini", "developer": "OpenAI", "license_type": "proprietary", "composite_iq_score": 95.8, "reasoning_score": 96.5, "agent_code_score": 68.4, "stem_gpqa_score": 84.1, "osworld_score": 48, "human_elo": 1329, "multimodal_score": 92.9, "context_window": "200k", "pricing_input_per_m": 1.1, "pricing_output_per_m": 4.4, "cache_read_per_m": 0.11, "reality_gap": 6.9, "category": "Efficiency & Edge", "key_moat": "Cost-efficient mathematical and competitive code generation", "release_date": "2026-06-15", "official_url": "https://openai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 13, "model_id": "openai-o1", "name": "OpenAI o1", "developer": "OpenAI", "license_type": "proprietary", "composite_iq_score": 96.5, "reasoning_score": 95.8, "agent_code_score": 71.2, "stem_gpqa_score": 86.4, "osworld_score": 52, "human_elo": 1339, "multimodal_score": 93.3, "context_window": "200k", "pricing_input_per_m": 15, "pricing_output_per_m": 60, "cache_read_per_m": 1.5, "reality_gap": 5.4, "category": "Frontier Reasoning", "key_moat": "Pioneering hidden chain-of-thought foundation system", "release_date": "2026-06-15", "official_url": "https://openai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 14, "model_id": "openai-o1-mini", "name": "OpenAI o1-mini", "developer": "OpenAI", "license_type": "proprietary", "composite_iq_score": 94.2, "reasoning_score": 94.9, "agent_code_score": 65.5, "stem_gpqa_score": 80.8, "osworld_score": 42, "human_elo": 1306, "multimodal_score": 92.1, "context_window": "128k", "pricing_input_per_m": 1.1, "pricing_output_per_m": 4.4, "cache_read_per_m": 0.11, "reality_gap": 5.9, "category": "Efficiency & Edge", "key_moat": "Specialized STEM inference accelerator", "release_date": "2026-06-15", "official_url": "https://openai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 15, "model_id": "gpt-4o", "name": "GPT-4o (Omni)", "developer": "OpenAI", "license_type": "proprietary", "composite_iq_score": 96, "reasoning_score": 96.2, "agent_code_score": 69.4, "stem_gpqa_score": 82, "osworld_score": 54, "human_elo": 1332, "multimodal_score": 95, "context_window": "128k", "pricing_input_per_m": 2.5, "pricing_output_per_m": 10, "cache_read_per_m": 0.25, "reality_gap": 5.8, "category": "Multimodal & Vision", "key_moat": "Real-time speech, vision, and text low-latency execution", "release_date": "2026-06-15", "official_url": "https://openai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 16, "model_id": "gpt-4o-mini", "name": "GPT-4o mini", "developer": "OpenAI", "license_type": "proprietary", "composite_iq_score": 93.4, "reasoning_score": 93.6, "agent_code_score": 60.2, "stem_gpqa_score": 75.6, "osworld_score": 38, "human_elo": 1294, "multimodal_score": 91.7, "context_window": "128k", "pricing_input_per_m": 0.15, "pricing_output_per_m": 0.6, "cache_read_per_m": 0.015, "reality_gap": 7.7, "category": "Efficiency & Edge", "key_moat": "Ubiquitous high-speed lightweight multi-modal engine", "release_date": "2026-06-15", "official_url": "https://openai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 17, "model_id": "gpt-4.5-preview", "name": "GPT-4.5 Preview", "developer": "OpenAI", "license_type": "proprietary", "composite_iq_score": 95.5, "reasoning_score": 95.7, "agent_code_score": 68, "stem_gpqa_score": 81.2, "osworld_score": 49, "human_elo": 1325, "multimodal_score": 92.8, "context_window": "128k", "pricing_input_per_m": 5, "pricing_output_per_m": 20, "cache_read_per_m": 0.5, "reality_gap": 7.3, "category": "Frontier Reasoning", "key_moat": "Large-scale world knowledge and nuanced stylometry", "release_date": "2026-06-15", "official_url": "https://openai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 18, "model_id": "gpt-4-turbo", "name": "GPT-4 Turbo", "developer": "OpenAI", "license_type": "proprietary", "composite_iq_score": 94.8, "reasoning_score": 94.2, "agent_code_score": 66.8, "stem_gpqa_score": 79.5, "osworld_score": 45, "human_elo": 1315, "multimodal_score": 92.4, "context_window": "128k", "pricing_input_per_m": 10, "pricing_output_per_m": 30, "cache_read_per_m": 1, "reality_gap": 9.6, "category": "Frontier Reasoning", "key_moat": "Established enterprise production standard", "release_date": "2026-06-15", "official_url": "https://openai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 19, "model_id": "gpt-3.5-turbo-0125", "name": "GPT-3.5 Turbo", "developer": "OpenAI", "license_type": "proprietary", "composite_iq_score": 86.2, "reasoning_score": 85.5, "agent_code_score": 44, "stem_gpqa_score": 58, "osworld_score": 20, "human_elo": 1190, "multimodal_score": 88.1, "context_window": "16k", "pricing_input_per_m": 0.5, "pricing_output_per_m": 1.5, "cache_read_per_m": 0.05, "reality_gap": 11.6, "category": "Efficiency & Edge", "key_moat": "Legacy fast triage and classification baseline", "release_date": "2026-06-15", "official_url": "https://openai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 20, "model_id": "claude-3-7-sonnet", "name": "Claude 3.7 Sonnet", "developer": "Anthropic", "license_type": "proprietary", "composite_iq_score": 97.5, "reasoning_score": 97.5, "agent_code_score": 70.3, "stem_gpqa_score": 84.8, "osworld_score": 58, "human_elo": 1354, "multimodal_score": 93.8, "context_window": "200k", "pricing_input_per_m": 3, "pricing_output_per_m": 15, "cache_read_per_m": 0.3, "reality_gap": 8.8, "category": "Coding & Agentic", "key_moat": "Dynamic thinking budget with industry-standard refactoring", "release_date": "2026-06-15", "official_url": "https://anthropic.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 21, "model_id": "claude-3-5-sonnet", "name": "Claude 3.5 Sonnet", "developer": "Anthropic", "license_type": "proprietary", "composite_iq_score": 96.2, "reasoning_score": 96.8, "agent_code_score": 67.2, "stem_gpqa_score": 82.5, "osworld_score": 53, "human_elo": 1335, "multimodal_score": 93.1, "context_window": "200k", "pricing_input_per_m": 3, "pricing_output_per_m": 15, "cache_read_per_m": 0.3, "reality_gap": 11.2, "category": "Coding & Agentic", "key_moat": "Desktop computer use automation pioneer", "release_date": "2026-06-15", "official_url": "https://anthropic.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 22, "model_id": "claude-3-5-haiku", "name": "Claude 3.5 Haiku", "developer": "Anthropic", "license_type": "proprietary", "composite_iq_score": 93.8, "reasoning_score": 93.4, "agent_code_score": 62, "stem_gpqa_score": 76.4, "osworld_score": 36, "human_elo": 1300, "multimodal_score": 91.9, "context_window": "200k", "pricing_input_per_m": 0.8, "pricing_output_per_m": 4, "cache_read_per_m": 0.08, "reality_gap": 9.5, "category": "Efficiency & Edge", "key_moat": "Ultra-fast low-latency code execution and extraction", "release_date": "2026-06-15", "official_url": "https://anthropic.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 23, "model_id": "claude-3-opus", "name": "Claude 3 Opus", "developer": "Anthropic", "license_type": "proprietary", "composite_iq_score": 94.5, "reasoning_score": 93.8, "agent_code_score": 64.1, "stem_gpqa_score": 79.2, "osworld_score": 40, "human_elo": 1310, "multimodal_score": 92.3, "context_window": "200k", "pricing_input_per_m": 15, "pricing_output_per_m": 75, "cache_read_per_m": 1.5, "reality_gap": 8.4, "category": "Frontier Reasoning", "key_moat": "Deep philosophical and complex prose synthesis", "release_date": "2026-06-15", "official_url": "https://anthropic.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 24, "model_id": "gemini-2-5-pro", "name": "Gemini 2.5 Pro", "developer": "Google DeepMind", "license_type": "proprietary", "composite_iq_score": 97, "reasoning_score": 97.1, "agent_code_score": 68.4, "stem_gpqa_score": 84, "osworld_score": 51, "human_elo": 1347, "multimodal_score": 95, "context_window": "2M", "pricing_input_per_m": 1.25, "pricing_output_per_m": 5, "cache_read_per_m": 0.125, "reality_gap": 7.6, "category": "Multimodal & Vision", "key_moat": "Deep research and long-context repository synthesis", "release_date": "2026-06-15", "official_url": "https://googledeepmind.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 25, "model_id": "gemini-2-flash", "name": "Gemini 2.0 Flash", "developer": "Google DeepMind", "license_type": "proprietary", "composite_iq_score": 94.6, "reasoning_score": 94.6, "agent_code_score": 63.8, "stem_gpqa_score": 78.5, "osworld_score": 41, "human_elo": 1312, "multimodal_score": 92.3, "context_window": "1M", "pricing_input_per_m": 0.1, "pricing_output_per_m": 0.4, "cache_read_per_m": 0.01, "reality_gap": 7.9, "category": "Efficiency & Edge", "key_moat": "Sub-second real-time streaming multimodal core", "release_date": "2026-06-15", "official_url": "https://googledeepmind.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 26, "model_id": "gemini-1-5-pro", "name": "Gemini 1.5 Pro", "developer": "Google DeepMind", "license_type": "proprietary", "composite_iq_score": 95, "reasoning_score": 95.1, "agent_code_score": 65, "stem_gpqa_score": 80, "osworld_score": 44, "human_elo": 1318, "multimodal_score": 95, "context_window": "2M", "pricing_input_per_m": 3.5, "pricing_output_per_m": 10.5, "cache_read_per_m": 0.35, "reality_gap": 6.9, "category": "Multimodal & Vision", "key_moat": "Breakthrough 2M token production context window", "release_date": "2026-06-15", "official_url": "https://googledeepmind.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 27, "model_id": "gemini-1-5-flash", "name": "Gemini 1.5 Flash", "developer": "Google DeepMind", "license_type": "proprietary", "composite_iq_score": 92.5, "reasoning_score": 91.9, "agent_code_score": 58.2, "stem_gpqa_score": 73, "osworld_score": 30, "human_elo": 1281, "multimodal_score": 91.3, "context_window": "1M", "pricing_input_per_m": 0.075, "pricing_output_per_m": 0.3, "cache_read_per_m": 0.007, "reality_gap": 5.5, "category": "Efficiency & Edge", "key_moat": "High-volume cost-optimized batch ingestion", "release_date": "2026-06-15", "official_url": "https://googledeepmind.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 28, "model_id": "gemma-2-27b", "name": "Gemma 2 27B", "developer": "Google DeepMind", "license_type": "open_weights", "composite_iq_score": 93, "reasoning_score": 93.6, "agent_code_score": 56.4, "stem_gpqa_score": 72.5, "osworld_score": 28, "human_elo": 1289, "multimodal_score": 91.5, "context_window": "8k", "pricing_input_per_m": 0.25, "pricing_output_per_m": 0.5, "cache_read_per_m": 0.025, "reality_gap": 11.2, "category": "Open-Weights SOTA", "key_moat": "Google premier open weights dense model", "release_date": "2026-06-15", "official_url": "https://googledeepmind.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 29, "model_id": "gemma-2-9b", "name": "Gemma 2 9B", "developer": "Google DeepMind", "license_type": "open_weights", "composite_iq_score": 90.4, "reasoning_score": 90.3, "agent_code_score": 50.1, "stem_gpqa_score": 66.8, "osworld_score": 22, "human_elo": 1251, "multimodal_score": 90.2, "context_window": "8k", "pricing_input_per_m": 0.1, "pricing_output_per_m": 0.2, "cache_read_per_m": 0.01, "reality_gap": 9.9, "category": "Open-Weights SOTA", "key_moat": "High efficiency edge-deployable open model", "release_date": "2026-06-15", "official_url": "https://googledeepmind.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 30, "model_id": "gemma-2-2b", "name": "Gemma 2 2B", "developer": "Google DeepMind", "license_type": "open_weights", "composite_iq_score": 84.5, "reasoning_score": 83.8, "agent_code_score": 38, "stem_gpqa_score": 54, "osworld_score": 14, "human_elo": 1165, "multimodal_score": 87.3, "context_window": "8k", "pricing_input_per_m": 0.05, "pricing_output_per_m": 0.1, "cache_read_per_m": 0.005, "reality_gap": 8.1, "category": "Efficiency & Edge", "key_moat": "On-device mobile and embedded AI engine", "release_date": "2026-06-15", "official_url": "https://googledeepmind.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 31, "model_id": "llama-4-behemoth-405b", "name": "Llama 4 Behemoth 405B", "developer": "Meta", "license_type": "open_weights", "composite_iq_score": 96.8, "reasoning_score": 97.1, "agent_code_score": 68.2, "stem_gpqa_score": 83.5, "osworld_score": 49, "human_elo": 1344, "multimodal_score": 93.4, "context_window": "128k", "pricing_input_per_m": 1.5, "pricing_output_per_m": 3.5, "cache_read_per_m": 0.15, "reality_gap": 6.6, "category": "Open-Weights SOTA", "key_moat": "Flagship sovereign open weights frontier model", "release_date": "2026-06-15", "official_url": "https://meta.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 32, "model_id": "llama-3-3-70b-instruct", "name": "Llama 3.3 70B Instruct", "developer": "Meta", "license_type": "open_weights", "composite_iq_score": 95.2, "reasoning_score": 94.6, "agent_code_score": 64.2, "stem_gpqa_score": 79.4, "osworld_score": 42, "human_elo": 1320, "multimodal_score": 92.6, "context_window": "128k", "pricing_input_per_m": 0.2, "pricing_output_per_m": 0.6, "cache_read_per_m": 0.02, "reality_gap": 10.9, "category": "Open-Weights SOTA", "key_moat": "Enterprise open-weights workhorse matching earlier 405B", "release_date": "2026-06-15", "official_url": "https://meta.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 33, "model_id": "llama-3-1-405b", "name": "Llama 3.1 405B", "developer": "Meta", "license_type": "open_weights", "composite_iq_score": 96, "reasoning_score": 95.3, "agent_code_score": 65.8, "stem_gpqa_score": 81.2, "osworld_score": 46, "human_elo": 1332, "multimodal_score": 93, "context_window": "128k", "pricing_input_per_m": 1.8, "pricing_output_per_m": 4, "cache_read_per_m": 0.18, "reality_gap": 11.1, "category": "Open-Weights SOTA", "key_moat": "Pioneering open frontier scale foundation weights", "release_date": "2026-06-15", "official_url": "https://meta.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 34, "model_id": "llama-3-1-70b", "name": "Llama 3.1 70B", "developer": "Meta", "license_type": "open_weights", "composite_iq_score": 94, "reasoning_score": 93.3, "agent_code_score": 61.5, "stem_gpqa_score": 77, "osworld_score": 38, "human_elo": 1303, "multimodal_score": 92, "context_window": "128k", "pricing_input_per_m": 0.25, "pricing_output_per_m": 0.7, "cache_read_per_m": 0.025, "reality_gap": 9.5, "category": "Open-Weights SOTA", "key_moat": "General-purpose high-performance open engine", "release_date": "2026-06-15", "official_url": "https://meta.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 35, "model_id": "llama-3-1-8b", "name": "Llama 3.1 8B", "developer": "Meta", "license_type": "open_weights", "composite_iq_score": 89.2, "reasoning_score": 88.8, "agent_code_score": 48.4, "stem_gpqa_score": 64, "osworld_score": 20, "human_elo": 1233, "multimodal_score": 89.6, "context_window": "128k", "pricing_input_per_m": 0.05, "pricing_output_per_m": 0.15, "cache_read_per_m": 0.005, "reality_gap": 11.1, "category": "Efficiency & Edge", "key_moat": "Lightweight local deployment standard", "release_date": "2026-06-15", "official_url": "https://meta.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 36, "model_id": "llama-3-2-90b-vision", "name": "Llama 3.2 90B Vision", "developer": "Meta", "license_type": "open_weights", "composite_iq_score": 94.8, "reasoning_score": 95.1, "agent_code_score": 62, "stem_gpqa_score": 77.5, "osworld_score": 44, "human_elo": 1315, "multimodal_score": 95, "context_window": "128k", "pricing_input_per_m": 0.4, "pricing_output_per_m": 0.9, "cache_read_per_m": 0.04, "reality_gap": 5, "category": "Multimodal & Vision", "key_moat": "Open multimodal document and image understanding", "release_date": "2026-06-15", "official_url": "https://meta.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 37, "model_id": "llama-3-2-11b-vision", "name": "Llama 3.2 11B Vision", "developer": "Meta", "license_type": "open_weights", "composite_iq_score": 91, "reasoning_score": 91.4, "agent_code_score": 52, "stem_gpqa_score": 68.2, "osworld_score": 26, "human_elo": 1260, "multimodal_score": 95, "context_window": "128k", "pricing_input_per_m": 0.08, "pricing_output_per_m": 0.2, "cache_read_per_m": 0.008, "reality_gap": 5, "category": "Multimodal & Vision", "key_moat": "Compact open-weights vision-language engine", "release_date": "2026-06-15", "official_url": "https://meta.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 38, "model_id": "llama-3-2-3b", "name": "Llama 3.2 3B", "developer": "Meta", "license_type": "open_weights", "composite_iq_score": 85, "reasoning_score": 84.3, "agent_code_score": 40.2, "stem_gpqa_score": 55.4, "osworld_score": 15, "human_elo": 1173, "multimodal_score": 87.5, "context_window": "128k", "pricing_input_per_m": 0.04, "pricing_output_per_m": 0.08, "cache_read_per_m": 0.004, "reality_gap": 4.5, "category": "Efficiency & Edge", "key_moat": "On-device mobile edge inference engine", "release_date": "2026-06-15", "official_url": "https://meta.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 39, "model_id": "llama-3-2-1b", "name": "Llama 3.2 1B", "developer": "Meta", "license_type": "open_weights", "composite_iq_score": 81.2, "reasoning_score": 81, "agent_code_score": 32, "stem_gpqa_score": 48, "osworld_score": 10, "human_elo": 1117, "multimodal_score": 85.6, "context_window": "128k", "pricing_input_per_m": 0.02, "pricing_output_per_m": 0.05, "cache_read_per_m": 0.002, "reality_gap": 5.3, "category": "Efficiency & Edge", "key_moat": "Embedded CPU-native ultra-light SLM", "release_date": "2026-06-15", "official_url": "https://meta.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 40, "model_id": "deepseek-r1", "name": "DeepSeek-R1", "developer": "DeepSeek AI", "license_type": "open_weights", "composite_iq_score": 97.8, "reasoning_score": 98, "agent_code_score": 67.9, "stem_gpqa_score": 87.2, "osworld_score": 48, "human_elo": 1358, "multimodal_score": 93.9, "context_window": "128k", "pricing_input_per_m": 0.55, "pricing_output_per_m": 2.19, "cache_read_per_m": 0.055, "reality_gap": 4.9, "category": "Frontier Reasoning", "key_moat": "Pure cold-start reinforcement learning frontier reasoning", "release_date": "2026-06-15", "official_url": "https://deepseekai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 41, "model_id": "deepseek-r1-distill-llama-70b", "name": "DeepSeek-R1-Distill-Llama-70B", "developer": "DeepSeek AI", "license_type": "open_weights", "composite_iq_score": 95, "reasoning_score": 94.5, "agent_code_score": 63.5, "stem_gpqa_score": 81, "osworld_score": 40, "human_elo": 1318, "multimodal_score": 92.5, "context_window": "128k", "pricing_input_per_m": 0.2, "pricing_output_per_m": 0.6, "cache_read_per_m": 0.02, "reality_gap": 5.5, "category": "Open-Weights SOTA", "key_moat": "Distilled dense reasoning power with Llama architecture", "release_date": "2026-06-15", "official_url": "https://deepseekai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 42, "model_id": "deepseek-r1-distill-qwen-32b", "name": "DeepSeek-R1-Distill-Qwen-32B", "developer": "DeepSeek AI", "license_type": "open_weights", "composite_iq_score": 94.6, "reasoning_score": 94.2, "agent_code_score": 64.8, "stem_gpqa_score": 80.4, "osworld_score": 38, "human_elo": 1312, "multimodal_score": 92.3, "context_window": "128k", "pricing_input_per_m": 0.15, "pricing_output_per_m": 0.45, "cache_read_per_m": 0.015, "reality_gap": 11.6, "category": "Coding & Agentic", "key_moat": "Coding and math specialist distilled on Qwen 32B", "release_date": "2026-06-15", "official_url": "https://deepseekai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 43, "model_id": "deepseek-r1-distill-qwen-14b", "name": "DeepSeek-R1-Distill-Qwen-14B", "developer": "DeepSeek AI", "license_type": "open_weights", "composite_iq_score": 92.4, "reasoning_score": 91.7, "agent_code_score": 56.2, "stem_gpqa_score": 74, "osworld_score": 28, "human_elo": 1280, "multimodal_score": 91.2, "context_window": "128k", "pricing_input_per_m": 0.1, "pricing_output_per_m": 0.3, "cache_read_per_m": 0.01, "reality_gap": 5.3, "category": "Efficiency & Edge", "key_moat": "High-density consumer GPU reasoning model", "release_date": "2026-06-15", "official_url": "https://deepseekai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 44, "model_id": "deepseek-r1-distill-llama-8b", "name": "DeepSeek-R1-Distill-Llama-8B", "developer": "DeepSeek AI", "license_type": "open_weights", "composite_iq_score": 89.8, "reasoning_score": 89.7, "agent_code_score": 49, "stem_gpqa_score": 66.5, "osworld_score": 20, "human_elo": 1242, "multimodal_score": 89.9, "context_window": "128k", "pricing_input_per_m": 0.05, "pricing_output_per_m": 0.15, "cache_read_per_m": 0.005, "reality_gap": 9.3, "category": "Efficiency & Edge", "key_moat": "Lightweight local reasoning model", "release_date": "2026-06-15", "official_url": "https://deepseekai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 45, "model_id": "deepseek-v3", "name": "DeepSeek-V3", "developer": "DeepSeek AI", "license_type": "open_weights", "composite_iq_score": 95.8, "reasoning_score": 96, "agent_code_score": 65.2, "stem_gpqa_score": 82, "osworld_score": 45, "human_elo": 1329, "multimodal_score": 92.9, "context_window": "128k", "pricing_input_per_m": 0.14, "pricing_output_per_m": 0.28, "cache_read_per_m": 0.014, "reality_gap": 7.7, "category": "Open-Weights SOTA", "key_moat": "671B Multi-head Latent Attention architecture with FP8 efficiency", "release_date": "2026-06-15", "official_url": "https://deepseekai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 46, "model_id": "deepseek-coder-v2-instruct", "name": "DeepSeek Coder V2 Instruct", "developer": "DeepSeek AI", "license_type": "open_weights", "composite_iq_score": 94.5, "reasoning_score": 94.3, "agent_code_score": 66.4, "stem_gpqa_score": 78, "osworld_score": 40, "human_elo": 1310, "multimodal_score": 92.3, "context_window": "128k", "pricing_input_per_m": 0.14, "pricing_output_per_m": 0.28, "cache_read_per_m": 0.014, "reality_gap": 6.3, "category": "Coding & Agentic", "key_moat": "Dedicated open coding model supporting 338 languages", "release_date": "2026-06-15", "official_url": "https://deepseekai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 47, "model_id": "qwen-2-5-max", "name": "Qwen 2.5 Max", "developer": "Alibaba Cloud", "license_type": "proprietary", "composite_iq_score": 96.5, "reasoning_score": 97.2, "agent_code_score": 66, "stem_gpqa_score": 82.8, "osworld_score": 48, "human_elo": 1339, "multimodal_score": 93.3, "context_window": "128k", "pricing_input_per_m": 1.6, "pricing_output_per_m": 6.4, "cache_read_per_m": 0.16, "reality_gap": 9.8, "category": "Frontier Reasoning", "key_moat": "Premier Asian multilingual reasoning engine", "release_date": "2026-06-15", "official_url": "https://alibabacloud.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 48, "model_id": "qwen-2-5-72b-instruct", "name": "Qwen 2.5 72B Instruct", "developer": "Alibaba Cloud", "license_type": "open_weights", "composite_iq_score": 95.4, "reasoning_score": 94.8, "agent_code_score": 64.5, "stem_gpqa_score": 80.2, "osworld_score": 44, "human_elo": 1323, "multimodal_score": 92.7, "context_window": "128k", "pricing_input_per_m": 0.25, "pricing_output_per_m": 0.75, "cache_read_per_m": 0.025, "reality_gap": 4.1, "category": "Open-Weights SOTA", "key_moat": "Flagship open weights model surpassing original Llama 3 70B", "release_date": "2026-06-15", "official_url": "https://alibabacloud.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 49, "model_id": "qwen-2-5-coder-32b-instruct", "name": "Qwen 2.5 Coder 32B", "developer": "Alibaba Cloud", "license_type": "open_weights", "composite_iq_score": 94.8, "reasoning_score": 94.8, "agent_code_score": 67.2, "stem_gpqa_score": 77, "osworld_score": 43, "human_elo": 1315, "multimodal_score": 92.4, "context_window": "128k", "pricing_input_per_m": 0.15, "pricing_output_per_m": 0.45, "cache_read_per_m": 0.015, "reality_gap": 5.4, "category": "Coding & Agentic", "key_moat": "Best-in-class 32B open code generation and agentic tool use", "release_date": "2026-06-15", "official_url": "https://alibabacloud.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 50, "model_id": "qwen-2-5-coder-14b-instruct", "name": "Qwen 2.5 Coder 14B", "developer": "Alibaba Cloud", "license_type": "open_weights", "composite_iq_score": 92.5, "reasoning_score": 91.9, "agent_code_score": 58, "stem_gpqa_score": 72, "osworld_score": 30, "human_elo": 1281, "multimodal_score": 91.3, "context_window": "128k", "pricing_input_per_m": 0.1, "pricing_output_per_m": 0.25, "cache_read_per_m": 0.01, "reality_gap": 9.7, "category": "Coding & Agentic", "key_moat": "Mid-size open coding specialist for developer workstations", "release_date": "2026-06-15", "official_url": "https://alibabacloud.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 51, "model_id": "qwen-2-5-coder-7b-instruct", "name": "Qwen 2.5 Coder 7B", "developer": "Alibaba Cloud", "license_type": "open_weights", "composite_iq_score": 89.6, "reasoning_score": 90.3, "agent_code_score": 49.5, "stem_gpqa_score": 64.2, "osworld_score": 22, "human_elo": 1239, "multimodal_score": 89.8, "context_window": "128k", "pricing_input_per_m": 0.06, "pricing_output_per_m": 0.15, "cache_read_per_m": 0.006, "reality_gap": 11.1, "category": "Efficiency & Edge", "key_moat": "Single GPU local agentic coding engine", "release_date": "2026-06-15", "official_url": "https://alibabacloud.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 52, "model_id": "qwen-2-5-32b-instruct", "name": "Qwen 2.5 32B Instruct", "developer": "Alibaba Cloud", "license_type": "open_weights", "composite_iq_score": 93.6, "reasoning_score": 93.2, "agent_code_score": 60.1, "stem_gpqa_score": 75.8, "osworld_score": 36, "human_elo": 1297, "multimodal_score": 91.8, "context_window": "128k", "pricing_input_per_m": 0.15, "pricing_output_per_m": 0.45, "cache_read_per_m": 0.015, "reality_gap": 7.5, "category": "Open-Weights SOTA", "key_moat": "Sweet-spot open-weights reasoning model", "release_date": "2026-06-15", "official_url": "https://alibabacloud.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 53, "model_id": "qwen-2-5-14b-instruct", "name": "Qwen 2.5 14B Instruct", "developer": "Alibaba Cloud", "license_type": "open_weights", "composite_iq_score": 91.5, "reasoning_score": 90.8, "agent_code_score": 54, "stem_gpqa_score": 70, "osworld_score": 26, "human_elo": 1267, "multimodal_score": 90.8, "context_window": "128k", "pricing_input_per_m": 0.08, "pricing_output_per_m": 0.2, "cache_read_per_m": 0.008, "reality_gap": 7.8, "category": "Efficiency & Edge", "key_moat": "Balanced bilingual general capability model", "release_date": "2026-06-15", "official_url": "https://alibabacloud.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 54, "model_id": "qwen-2-5-7b-instruct", "name": "Qwen 2.5 7B Instruct", "developer": "Alibaba Cloud", "license_type": "open_weights", "composite_iq_score": 88, "reasoning_score": 88.5, "agent_code_score": 46.2, "stem_gpqa_score": 62, "osworld_score": 18, "human_elo": 1216, "multimodal_score": 89, "context_window": "128k", "pricing_input_per_m": 0.05, "pricing_output_per_m": 0.12, "cache_read_per_m": 0.005, "reality_gap": 5, "category": "Efficiency & Edge", "key_moat": "Highly capable edge model for routine classification", "release_date": "2026-06-15", "official_url": "https://alibabacloud.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 55, "model_id": "qwen-2-vl-72b-instruct", "name": "Qwen 2 VL 72B", "developer": "Alibaba Cloud", "license_type": "open_weights", "composite_iq_score": 94.6, "reasoning_score": 95.1, "agent_code_score": 59, "stem_gpqa_score": 78, "osworld_score": 45, "human_elo": 1312, "multimodal_score": 95, "context_window": "128k", "pricing_input_per_m": 0.4, "pricing_output_per_m": 1, "cache_read_per_m": 0.04, "reality_gap": 9.5, "category": "Multimodal & Vision", "key_moat": "Frontier open vision-language model with video understanding", "release_date": "2026-06-15", "official_url": "https://alibabacloud.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 56, "model_id": "qwen-2-vl-7b-instruct", "name": "Qwen 2 VL 7B", "developer": "Alibaba Cloud", "license_type": "open_weights", "composite_iq_score": 90.2, "reasoning_score": 89.6, "agent_code_score": 50.4, "stem_gpqa_score": 67, "osworld_score": 28, "human_elo": 1248, "multimodal_score": 95, "context_window": "128k", "pricing_input_per_m": 0.1, "pricing_output_per_m": 0.25, "cache_read_per_m": 0.01, "reality_gap": 6.5, "category": "Multimodal & Vision", "key_moat": "Efficient edge vision model for mobile and IoT applications", "release_date": "2026-06-15", "official_url": "https://alibabacloud.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 57, "model_id": "mistral-large-2", "name": "Mistral Large 2", "developer": "Mistral AI", "license_type": "proprietary", "composite_iq_score": 95, "reasoning_score": 94.3, "agent_code_score": 64, "stem_gpqa_score": 80.5, "osworld_score": 46, "human_elo": 1318, "multimodal_score": 92.5, "context_window": "128k", "pricing_input_per_m": 2, "pricing_output_per_m": 6, "cache_read_per_m": 0.2, "reality_gap": 5.4, "category": "Frontier Reasoning", "key_moat": "European frontier flagship with 80+ language multilingual support", "release_date": "2026-06-15", "official_url": "https://mistralai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 58, "model_id": "pixtral-large", "name": "Pixtral Large", "developer": "Mistral AI", "license_type": "proprietary", "composite_iq_score": 94.8, "reasoning_score": 94.5, "agent_code_score": 62.5, "stem_gpqa_score": 78.4, "osworld_score": 44, "human_elo": 1315, "multimodal_score": 95, "context_window": "128k", "pricing_input_per_m": 2, "pricing_output_per_m": 6, "cache_read_per_m": 0.2, "reality_gap": 6.6, "category": "Multimodal & Vision", "key_moat": "124B frontier multimodal image & document engine", "release_date": "2026-06-15", "official_url": "https://mistralai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 59, "model_id": "codestral-2501", "name": "Codestral 25.01", "developer": "Mistral AI", "license_type": "proprietary", "composite_iq_score": 94.2, "reasoning_score": 93.7, "agent_code_score": 65.8, "stem_gpqa_score": 76.5, "osworld_score": 42, "human_elo": 1306, "multimodal_score": 92.1, "context_window": "256k", "pricing_input_per_m": 0.3, "pricing_output_per_m": 0.9, "cache_read_per_m": 0.03, "reality_gap": 9.6, "category": "Coding & Agentic", "key_moat": "Fast 256k context dedicated coding engine", "release_date": "2026-06-15", "official_url": "https://mistralai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 60, "model_id": "mistral-embed", "name": "Mistral Embed", "developer": "Mistral AI", "license_type": "proprietary", "composite_iq_score": 88, "reasoning_score": 87.6, "agent_code_score": 30, "stem_gpqa_score": 50, "osworld_score": 10, "human_elo": 1216, "multimodal_score": 89, "context_window": "8k", "pricing_input_per_m": 0.1, "pricing_output_per_m": 0.1, "cache_read_per_m": 0.01, "reality_gap": 6.2, "category": "Efficiency & Edge", "key_moat": "Dense semantic vector retrieval embedding model", "release_date": "2026-06-15", "official_url": "https://mistralai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 61, "model_id": "mistral-nemo-12b", "name": "Mistral NeMo 12B", "developer": "Mistral AI", "license_type": "open_weights", "composite_iq_score": 91.2, "reasoning_score": 91.1, "agent_code_score": 52, "stem_gpqa_score": 68, "osworld_score": 25, "human_elo": 1262, "multimodal_score": 90.6, "context_window": "128k", "pricing_input_per_m": 0.15, "pricing_output_per_m": 0.15, "cache_read_per_m": 0.015, "reality_gap": 5.5, "category": "Open-Weights SOTA", "key_moat": "NVIDIA co-developed open weights efficient model", "release_date": "2026-06-15", "official_url": "https://mistralai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 62, "model_id": "ministral-8b", "name": "Ministral 8B", "developer": "Mistral AI", "license_type": "proprietary", "composite_iq_score": 89.5, "reasoning_score": 90, "agent_code_score": 48, "stem_gpqa_score": 63.5, "osworld_score": 20, "human_elo": 1238, "multimodal_score": 89.8, "context_window": "128k", "pricing_input_per_m": 0.1, "pricing_output_per_m": 0.1, "cache_read_per_m": 0.01, "reality_gap": 5.3, "category": "Efficiency & Edge", "key_moat": "Compact edge model optimized for low-latency reasoning", "release_date": "2026-06-15", "official_url": "https://mistralai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 63, "model_id": "ministral-3b", "name": "Ministral 3B", "developer": "Mistral AI", "license_type": "proprietary", "composite_iq_score": 85.4, "reasoning_score": 85.4, "agent_code_score": 41.2, "stem_gpqa_score": 56, "osworld_score": 15, "human_elo": 1178, "multimodal_score": 87.7, "context_window": "128k", "pricing_input_per_m": 0.04, "pricing_output_per_m": 0.04, "cache_read_per_m": 0.004, "reality_gap": 8.8, "category": "Efficiency & Edge", "key_moat": "Sub-4B ultra-compact edge model", "release_date": "2026-06-15", "official_url": "https://mistralai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 64, "model_id": "pixtral-12b", "name": "Pixtral 12B", "developer": "Mistral AI", "license_type": "open_weights", "composite_iq_score": 90.5, "reasoning_score": 90.2, "agent_code_score": 50, "stem_gpqa_score": 66.8, "osworld_score": 27, "human_elo": 1252, "multimodal_score": 95, "context_window": "128k", "pricing_input_per_m": 0.15, "pricing_output_per_m": 0.15, "cache_read_per_m": 0.015, "reality_gap": 8.8, "category": "Multimodal & Vision", "key_moat": "Open weights vision-language model", "release_date": "2026-06-15", "official_url": "https://mistralai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 65, "model_id": "grok-3-ultra", "name": "Grok 3 Ultra", "developer": "xAI", "license_type": "proprietary", "composite_iq_score": 96.5, "reasoning_score": 96.4, "agent_code_score": 67.5, "stem_gpqa_score": 83, "osworld_score": 50, "human_elo": 1339, "multimodal_score": 93.3, "context_window": "1M", "pricing_input_per_m": 4, "pricing_output_per_m": 12, "cache_read_per_m": 0.4, "reality_gap": 6.1, "category": "Frontier Reasoning", "key_moat": "Colossus 100k H100 pre-trained frontier reasoning model", "release_date": "2026-06-15", "official_url": "https://xai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 66, "model_id": "grok-2", "name": "Grok 2", "developer": "xAI", "license_type": "proprietary", "composite_iq_score": 94.5, "reasoning_score": 94.2, "agent_code_score": 62, "stem_gpqa_score": 78, "osworld_score": 42, "human_elo": 1310, "multimodal_score": 92.3, "context_window": "128k", "pricing_input_per_m": 2, "pricing_output_per_m": 10, "cache_read_per_m": 0.2, "reality_gap": 6.3, "category": "Frontier Reasoning", "key_moat": "Real-time X telemetry reasoning engine", "release_date": "2026-06-15", "official_url": "https://xai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 67, "model_id": "grok-2-mini", "name": "Grok 2 Mini", "developer": "xAI", "license_type": "proprietary", "composite_iq_score": 92, "reasoning_score": 92.3, "agent_code_score": 54, "stem_gpqa_score": 71, "osworld_score": 28, "human_elo": 1274, "multimodal_score": 91, "context_window": "128k", "pricing_input_per_m": 0.2, "pricing_output_per_m": 1, "cache_read_per_m": 0.02, "reality_gap": 6.4, "category": "Efficiency & Edge", "key_moat": "Low-latency cost-effective Grok variant", "release_date": "2026-06-15", "official_url": "https://xai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 68, "model_id": "phi-4-14b", "name": "Phi-4 14B", "developer": "Microsoft", "license_type": "open_weights", "composite_iq_score": 93.8, "reasoning_score": 94, "agent_code_score": 62.4, "stem_gpqa_score": 78.5, "osworld_score": 36, "human_elo": 1300, "multimodal_score": 91.9, "context_window": "16k", "pricing_input_per_m": 0.12, "pricing_output_per_m": 0.35, "cache_read_per_m": 0.012, "reality_gap": 7.4, "category": "Open-Weights SOTA", "key_moat": "Synthetic textbook pre-training reasoning density champion", "release_date": "2026-06-15", "official_url": "https://microsoft.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 69, "model_id": "phi-3-5-moe-instruct", "name": "Phi-3.5 MoE Instruct", "developer": "Microsoft", "license_type": "open_weights", "composite_iq_score": 92, "reasoning_score": 92.7, "agent_code_score": 56, "stem_gpqa_score": 73, "osworld_score": 30, "human_elo": 1274, "multimodal_score": 91, "context_window": "128k", "pricing_input_per_m": 0.1, "pricing_output_per_m": 0.3, "cache_read_per_m": 0.01, "reality_gap": 9, "category": "Efficiency & Edge", "key_moat": "16x3.8B sparse MoE activating only 6.6B parameters", "release_date": "2026-06-15", "official_url": "https://microsoft.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 70, "model_id": "phi-3-5-mini-instruct", "name": "Phi-3.5 Mini Instruct", "developer": "Microsoft", "license_type": "open_weights", "composite_iq_score": 88.5, "reasoning_score": 88.1, "agent_code_score": 47, "stem_gpqa_score": 64, "osworld_score": 19, "human_elo": 1223, "multimodal_score": 89.3, "context_window": "128k", "pricing_input_per_m": 0.05, "pricing_output_per_m": 0.15, "cache_read_per_m": 0.005, "reality_gap": 10.3, "category": "Efficiency & Edge", "key_moat": "3.8B parameter edge execution benchmark", "release_date": "2026-06-15", "official_url": "https://microsoft.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 71, "model_id": "phi-3-5-vision-instruct", "name": "Phi-3.5 Vision", "developer": "Microsoft", "license_type": "open_weights", "composite_iq_score": 89.8, "reasoning_score": 90.4, "agent_code_score": 48, "stem_gpqa_score": 66, "osworld_score": 24, "human_elo": 1242, "multimodal_score": 95, "context_window": "128k", "pricing_input_per_m": 0.08, "pricing_output_per_m": 0.24, "cache_read_per_m": 0.008, "reality_gap": 6.4, "category": "Multimodal & Vision", "key_moat": "Multimodal lightweight chart, diagram, and image analyst", "release_date": "2026-06-15", "official_url": "https://microsoft.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 72, "model_id": "command-r-plus-08-2024", "name": "Command R+ (Aug 2024)", "developer": "Cohere", "license_type": "proprietary", "composite_iq_score": 94.2, "reasoning_score": 94.9, "agent_code_score": 61, "stem_gpqa_score": 76.8, "osworld_score": 39, "human_elo": 1306, "multimodal_score": 92.1, "context_window": "128k", "pricing_input_per_m": 2.5, "pricing_output_per_m": 10, "cache_read_per_m": 0.25, "reality_gap": 6.8, "category": "Enterprise & Math", "key_moat": "Enterprise RAG and automated tool use specialist", "release_date": "2026-06-15", "official_url": "https://cohere.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 73, "model_id": "command-r-08-2024", "name": "Command R (Aug 2024)", "developer": "Cohere", "license_type": "proprietary", "composite_iq_score": 91.5, "reasoning_score": 91.1, "agent_code_score": 53, "stem_gpqa_score": 70, "osworld_score": 27, "human_elo": 1267, "multimodal_score": 90.8, "context_window": "128k", "pricing_input_per_m": 0.15, "pricing_output_per_m": 0.6, "cache_read_per_m": 0.015, "reality_gap": 9, "category": "Efficiency & Edge", "key_moat": "Fast enterprise retrieval-augmented generation engine", "release_date": "2026-06-15", "official_url": "https://cohere.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 74, "model_id": "command-r7b-12-2024", "name": "Command R7B", "developer": "Cohere", "license_type": "open_weights", "composite_iq_score": 89, "reasoning_score": 89.3, "agent_code_score": 47.5, "stem_gpqa_score": 63, "osworld_score": 21, "human_elo": 1231, "multimodal_score": 89.5, "context_window": "128k", "pricing_input_per_m": 0.037, "pricing_output_per_m": 0.15, "cache_read_per_m": 0.004, "reality_gap": 10.1, "category": "Open-Weights SOTA", "key_moat": "Open weights multilingual RAG specialist", "release_date": "2026-06-15", "official_url": "https://cohere.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 75, "model_id": "embed-v3", "name": "Cohere Embed v3.0", "developer": "Cohere", "license_type": "proprietary", "composite_iq_score": 88.5, "reasoning_score": 88.1, "agent_code_score": 25, "stem_gpqa_score": 50, "osworld_score": 10, "human_elo": 1223, "multimodal_score": 89.3, "context_window": "512", "pricing_input_per_m": 0.1, "pricing_output_per_m": 0.1, "cache_read_per_m": 0.01, "reality_gap": 5.4, "category": "Efficiency & Edge", "key_moat": "Industry-standard search and retrieval embedding vectorizer", "release_date": "2026-06-15", "official_url": "https://cohere.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 76, "model_id": "glm-4-plus", "name": "GLM-4-Plus", "developer": "Zhipu AI", "license_type": "proprietary", "composite_iq_score": 94.2, "reasoning_score": 94.7, "agent_code_score": 63, "stem_gpqa_score": 78, "osworld_score": 41, "human_elo": 1319, "multimodal_score": 84, "context_window": "128k", "pricing_input_per_m": 1.4, "pricing_output_per_m": 4.2, "cache_read_per_m": 0.21, "reality_gap": 10.2, "category": "Frontier Reasoning", "key_moat": "Bilingual Chinese tool-use orchestration", "release_date": "2026-05-20", "official_url": "https://zhipuai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 77, "model_id": "glm-4-air", "name": "GLM-4-Air", "developer": "Zhipu AI", "license_type": "proprietary", "composite_iq_score": 91, "reasoning_score": 91.5, "agent_code_score": 54, "stem_gpqa_score": 70, "osworld_score": 28, "human_elo": 1274, "multimodal_score": 84, "context_window": "128k", "pricing_input_per_m": 0.14, "pricing_output_per_m": 0.14, "cache_read_per_m": 0.021, "reality_gap": 10.1, "category": "Efficiency & Edge", "key_moat": "High-speed cost-effective enterprise agent", "release_date": "2026-05-20", "official_url": "https://zhipuai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 78, "model_id": "glm-4-9b-chat", "name": "GLM-4-9B-Chat", "developer": "Zhipu AI", "license_type": "open_weights", "composite_iq_score": 89.8, "reasoning_score": 90.3, "agent_code_score": 50, "stem_gpqa_score": 66, "osworld_score": 24, "human_elo": 1257, "multimodal_score": 84, "context_window": "128k", "pricing_input_per_m": 0.1, "pricing_output_per_m": 0.1, "cache_read_per_m": 0.015, "reality_gap": 6.1, "category": "Open-Weights SOTA", "key_moat": "Open-weights bilingual 9B transformer", "release_date": "2026-05-20", "official_url": "https://zhipuai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 79, "model_id": "yi-lightning", "name": "Yi-Lightning", "developer": "01.AI", "license_type": "proprietary", "composite_iq_score": 94.5, "reasoning_score": 95, "agent_code_score": 63.8, "stem_gpqa_score": 78.5, "osworld_score": 40, "human_elo": 1323, "multimodal_score": 84, "context_window": "16k", "pricing_input_per_m": 0.14, "pricing_output_per_m": 0.14, "cache_read_per_m": 0.021, "reality_gap": 11.4, "category": "Efficiency & Edge", "key_moat": "Sub-100ms ultra-fast inference reasoning engine", "release_date": "2026-05-20", "official_url": "https://01ai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 80, "model_id": "yi-large", "name": "Yi-Large", "developer": "01.AI", "license_type": "proprietary", "composite_iq_score": 93.8, "reasoning_score": 94.3, "agent_code_score": 60.5, "stem_gpqa_score": 76, "osworld_score": 38, "human_elo": 1313, "multimodal_score": 84, "context_window": "32k", "pricing_input_per_m": 3, "pricing_output_per_m": 3, "cache_read_per_m": 0.45, "reality_gap": 8.5, "category": "Frontier Reasoning", "key_moat": "Comprehensive bilingual Chinese-English model", "release_date": "2026-05-20", "official_url": "https://01ai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 81, "model_id": "yi-1-5-34b-chat", "name": "Yi-1.5-34B-Chat", "developer": "01.AI", "license_type": "open_weights", "composite_iq_score": 92, "reasoning_score": 92.5, "agent_code_score": 56, "stem_gpqa_score": 71, "osworld_score": 30, "human_elo": 1288, "multimodal_score": 84, "context_window": "32k", "pricing_input_per_m": 0.2, "pricing_output_per_m": 0.2, "cache_read_per_m": 0.03, "reality_gap": 8.7, "category": "Open-Weights SOTA", "key_moat": "Open weights dense 34B architecture", "release_date": "2026-05-20", "official_url": "https://01ai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 82, "model_id": "baichuan-4", "name": "Baichuan-4", "developer": "Baichuan AI", "license_type": "proprietary", "composite_iq_score": 93, "reasoning_score": 93.5, "agent_code_score": 58, "stem_gpqa_score": 74.5, "osworld_score": 35, "human_elo": 1302, "multimodal_score": 84, "context_window": "128k", "pricing_input_per_m": 1.5, "pricing_output_per_m": 4.5, "cache_read_per_m": 0.225, "reality_gap": 11.1, "category": "Enterprise & Math", "key_moat": "Medical and enterprise Chinese domain specialist", "release_date": "2026-05-20", "official_url": "https://baichuanai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 83, "model_id": "sensechat-5-5", "name": "SenseChat-5.5", "developer": "SenseTime", "license_type": "proprietary", "composite_iq_score": 92.5, "reasoning_score": 93, "agent_code_score": 57, "stem_gpqa_score": 73, "osworld_score": 34, "human_elo": 1295, "multimodal_score": 94, "context_window": "128k", "pricing_input_per_m": 1, "pricing_output_per_m": 3, "cache_read_per_m": 0.15, "reality_gap": 11.4, "category": "Multimodal & Vision", "key_moat": "Visual perception and industrial automation", "release_date": "2026-05-20", "official_url": "https://sensetime.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 84, "model_id": "hunyuan-large", "name": "Hunyuan-Large", "developer": "Tencent", "license_type": "open_weights", "composite_iq_score": 94, "reasoning_score": 94.5, "agent_code_score": 61.2, "stem_gpqa_score": 77, "osworld_score": 39, "human_elo": 1316, "multimodal_score": 84, "context_window": "256k", "pricing_input_per_m": 0.5, "pricing_output_per_m": 1.5, "cache_read_per_m": 0.075, "reality_gap": 6.1, "category": "Open-Weights SOTA", "key_moat": "389B MoE with 52B active parameters", "release_date": "2026-05-20", "official_url": "https://tencent.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 85, "model_id": "step-2-16k", "name": "Step-2-16k", "developer": "StepFun", "license_type": "proprietary", "composite_iq_score": 93.2, "reasoning_score": 93.7, "agent_code_score": 59, "stem_gpqa_score": 75, "osworld_score": 36, "human_elo": 1305, "multimodal_score": 94, "context_window": "16k", "pricing_input_per_m": 2, "pricing_output_per_m": 5, "cache_read_per_m": 0.3, "reality_gap": 5.8, "category": "Multimodal & Vision", "key_moat": "Complex multimodal chart and reasoning specialist", "release_date": "2026-05-20", "official_url": "https://stepfun.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 86, "model_id": "minimax-abab-6-5", "name": "MiniMax-ABAB-6.5", "developer": "MiniMax", "license_type": "proprietary", "composite_iq_score": 92.8, "reasoning_score": 93.3, "agent_code_score": 57.5, "stem_gpqa_score": 74, "osworld_score": 33, "human_elo": 1299, "multimodal_score": 84, "context_window": "245k", "pricing_input_per_m": 1, "pricing_output_per_m": 1, "cache_read_per_m": 0.15, "reality_gap": 11.5, "category": "Efficiency & Edge", "key_moat": "Linear attention MoE processing 245k tokens", "release_date": "2026-05-20", "official_url": "https://minimax.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 87, "model_id": "cursor-small", "name": "Cursor-Small", "developer": "Anysphere", "license_type": "proprietary", "composite_iq_score": 93.5, "reasoning_score": 94, "agent_code_score": 66.5, "stem_gpqa_score": 74, "osworld_score": 48, "human_elo": 1309, "multimodal_score": 84, "context_window": "32k", "pricing_input_per_m": 0.2, "pricing_output_per_m": 0.8, "cache_read_per_m": 0.03, "reality_gap": 10.6, "category": "Coding & Agentic", "key_moat": "Next-action and surgical diff generation engine", "release_date": "2026-05-20", "official_url": "https://anysphere.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 88, "model_id": "starcoder-2-15b", "name": "StarCoder-2-15B", "developer": "BigCode", "license_type": "open_weights", "composite_iq_score": 89.2, "reasoning_score": 89.7, "agent_code_score": 49, "stem_gpqa_score": 64, "osworld_score": 22, "human_elo": 1249, "multimodal_score": 84, "context_window": "16k", "pricing_input_per_m": 0.1, "pricing_output_per_m": 0.2, "cache_read_per_m": 0.015, "reality_gap": 10, "category": "Coding & Agentic", "key_moat": "Transparent open data code pre-training", "release_date": "2026-05-20", "official_url": "https://bigcode.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 89, "model_id": "codegemma-7b", "name": "CodeGemma-7B", "developer": "Google DeepMind", "license_type": "open_weights", "composite_iq_score": 88.5, "reasoning_score": 89, "agent_code_score": 47.2, "stem_gpqa_score": 62, "osworld_score": 20, "human_elo": 1239, "multimodal_score": 84, "context_window": "8k", "pricing_input_per_m": 0.08, "pricing_output_per_m": 0.15, "cache_read_per_m": 0.012, "reality_gap": 9.4, "category": "Coding & Agentic", "key_moat": "Code completion and infilling specialist", "release_date": "2026-05-20", "official_url": "https://googledeepmind.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 90, "model_id": "granite-3-8b-code", "name": "Granite-3-8B-Code", "developer": "IBM", "license_type": "open_weights", "composite_iq_score": 89, "reasoning_score": 89.5, "agent_code_score": 48, "stem_gpqa_score": 63, "osworld_score": 21, "human_elo": 1246, "multimodal_score": 84, "context_window": "128k", "pricing_input_per_m": 0.08, "pricing_output_per_m": 0.16, "cache_read_per_m": 0.012, "reality_gap": 11.8, "category": "Coding & Agentic", "key_moat": "Enterprise Apache 2.0 indemnified code model", "release_date": "2026-05-20", "official_url": "https://ibm.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 91, "model_id": "granite-3-8b-instruct", "name": "Granite-3-8B-Instruct", "developer": "IBM", "license_type": "open_weights", "composite_iq_score": 89.5, "reasoning_score": 90, "agent_code_score": 47, "stem_gpqa_score": 64.5, "osworld_score": 20, "human_elo": 1253, "multimodal_score": 84, "context_window": "128k", "pricing_input_per_m": 0.08, "pricing_output_per_m": 0.16, "cache_read_per_m": 0.012, "reality_gap": 7.6, "category": "Enterprise & Math", "key_moat": "Enterprise compliant governance and tool-calling", "release_date": "2026-05-20", "official_url": "https://ibm.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 92, "model_id": "reflection-70b", "name": "Reflection-70B", "developer": "HyperWrite", "license_type": "open_weights", "composite_iq_score": 93, "reasoning_score": 93.5, "agent_code_score": 59, "stem_gpqa_score": 76, "osworld_score": 37, "human_elo": 1302, "multimodal_score": 84, "context_window": "128k", "pricing_input_per_m": 0.3, "pricing_output_per_m": 0.9, "cache_read_per_m": 0.045, "reality_gap": 9, "category": "Frontier Reasoning", "key_moat": "Reflection tuning error-correction paradigm", "release_date": "2026-05-20", "official_url": "https://hyperwrite.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 93, "model_id": "hermes-3-llama-3-1-70b", "name": "Hermes-3-Llama-3-1-70B", "developer": "Nous Research", "license_type": "open_weights", "composite_iq_score": 94.4, "reasoning_score": 94.9, "agent_code_score": 62.8, "stem_gpqa_score": 78.5, "osworld_score": 41, "human_elo": 1322, "multimodal_score": 84, "context_window": "128k", "pricing_input_per_m": 0.3, "pricing_output_per_m": 0.8, "cache_read_per_m": 0.045, "reality_gap": 6.9, "category": "Coding & Agentic", "key_moat": "Uncensored multi-turn agentic orchestration", "release_date": "2026-05-20", "official_url": "https://nousresearch.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 94, "model_id": "hermes-3-llama-3-1-8b", "name": "Hermes-3-Llama-3-1-8B", "developer": "Nous Research", "license_type": "open_weights", "composite_iq_score": 89.5, "reasoning_score": 90, "agent_code_score": 48.5, "stem_gpqa_score": 64.5, "osworld_score": 21, "human_elo": 1253, "multimodal_score": 84, "context_window": "128k", "pricing_input_per_m": 0.06, "pricing_output_per_m": 0.15, "cache_read_per_m": 0.009, "reality_gap": 6.4, "category": "Efficiency & Edge", "key_moat": "Compact function calling and roleplay engine", "release_date": "2026-05-20", "official_url": "https://nousresearch.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 95, "model_id": "aria-multimodal", "name": "Aria-Multimodal", "developer": "Rhymes AI", "license_type": "open_weights", "composite_iq_score": 93.8, "reasoning_score": 94.3, "agent_code_score": 58, "stem_gpqa_score": 76, "osworld_score": 42, "human_elo": 1313, "multimodal_score": 94, "context_window": "64k", "pricing_input_per_m": 0.25, "pricing_output_per_m": 0.5, "cache_read_per_m": 0.037, "reality_gap": 6.2, "category": "Multimodal & Vision", "key_moat": "Native MoE multimodal video and web agent", "release_date": "2026-05-20", "official_url": "https://rhymesai.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 96, "model_id": "smollm-2-1-7b", "name": "SmolLM-2-1.7B", "developer": "Hugging Face", "license_type": "open_weights", "composite_iq_score": 82, "reasoning_score": 82.5, "agent_code_score": 34, "stem_gpqa_score": 50, "osworld_score": 12, "human_elo": 1148, "multimodal_score": 84, "context_window": "8k", "pricing_input_per_m": 0.02, "pricing_output_per_m": 0.04, "cache_read_per_m": 0.003, "reality_gap": 7.6, "category": "Efficiency & Edge", "key_moat": "Ultra-compact edge dataset distillation", "release_date": "2026-05-20", "official_url": "https://huggingface.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 97, "model_id": "smollm-2-360m", "name": "SmolLM-2-360M", "developer": "Hugging Face", "license_type": "open_weights", "composite_iq_score": 76, "reasoning_score": 76.5, "agent_code_score": 24, "stem_gpqa_score": 42, "osworld_score": 6, "human_elo": 1064, "multimodal_score": 84, "context_window": "8k", "pricing_input_per_m": 0.01, "pricing_output_per_m": 0.02, "cache_read_per_m": 0.002, "reality_gap": 11.5, "category": "Efficiency & Edge", "key_moat": "Sub-500M browser web assembly execution", "release_date": "2026-05-20", "official_url": "https://huggingface.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 98, "model_id": "llama-guard-3-8b", "name": "Llama-Guard-3-8B", "developer": "Meta", "license_type": "open_weights", "composite_iq_score": 88, "reasoning_score": 88.5, "agent_code_score": 35, "stem_gpqa_score": 58, "osworld_score": 15, "human_elo": 1232, "multimodal_score": 84, "context_window": "128k", "pricing_input_per_m": 0.05, "pricing_output_per_m": 0.1, "cache_read_per_m": 0.007, "reality_gap": 6.1, "category": "Enterprise & Math", "key_moat": "Safety taxonomy moderation and hazard classifier", "release_date": "2026-05-20", "official_url": "https://meta.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 99, "model_id": "nemotron-4-340b-instruct", "name": "Nemotron-4-340B-Instruct", "developer": "NVIDIA", "license_type": "open_weights", "composite_iq_score": 95.5, "reasoning_score": 96, "agent_code_score": 65, "stem_gpqa_score": 81, "osworld_score": 45, "human_elo": 1337, "multimodal_score": 84, "context_window": "128k", "pricing_input_per_m": 1, "pricing_output_per_m": 2.5, "cache_read_per_m": 0.15, "reality_gap": 6.3, "category": "Open-Weights SOTA", "key_moat": "Synthetic data generation and fine-tuning engine", "release_date": "2026-05-20", "official_url": "https://nvidia.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 100, "model_id": "nemotron-4-340b-reward", "name": "Nemotron-4-340B-Reward", "developer": "NVIDIA", "license_type": "open_weights", "composite_iq_score": 94.8, "reasoning_score": 95.3, "agent_code_score": 60, "stem_gpqa_score": 79, "osworld_score": 40, "human_elo": 1327, "multimodal_score": 84, "context_window": "128k", "pricing_input_per_m": 1, "pricing_output_per_m": 2.5, "cache_read_per_m": 0.15, "reality_gap": 7.4, "category": "Enterprise & Math", "key_moat": "High-dimensional RLHF reward modeling standard", "release_date": "2026-05-20", "official_url": "https://nvidia.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 101, "model_id": "llama-3-groq-tool-use-70b", "name": "Llama-3-Groq-Tool-Use-70B", "developer": "Groq", "license_type": "open_weights", "composite_iq_score": 94, "reasoning_score": 94.5, "agent_code_score": 63.5, "stem_gpqa_score": 77, "osworld_score": 41, "human_elo": 1316, "multimodal_score": 84, "context_window": "8k", "pricing_input_per_m": 0.3, "pricing_output_per_m": 0.9, "cache_read_per_m": 0.045, "reality_gap": 5.8, "category": "Coding & Agentic", "key_moat": "LPU hardware-accelerated tool calling execution", "release_date": "2026-05-20", "official_url": "https://groq.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 102, "model_id": "llama-3-groq-tool-use-8b", "name": "Llama-3-Groq-Tool-Use-8B", "developer": "Groq", "license_type": "open_weights", "composite_iq_score": 89, "reasoning_score": 89.5, "agent_code_score": 50, "stem_gpqa_score": 64, "osworld_score": 24, "human_elo": 1246, "multimodal_score": 84, "context_window": "8k", "pricing_input_per_m": 0.08, "pricing_output_per_m": 0.2, "cache_read_per_m": 0.012, "reality_gap": 7.1, "category": "Efficiency & Edge", "key_moat": "Sub-50ms tool triage on Groq LPUs", "release_date": "2026-05-20", "official_url": "https://groq.com", "status": "verified", "evaluation_tier": "empirical" }, { "rank": 103, "model_id": "modl-salesforce-v1-103", "name": "Salesforce Model-103 (Coding & Agentic)", "developer": "Salesforce", "license_type": "open_weights", "composite_iq_score": 88.5, "reasoning_score": 88.2, "agent_code_score": 53.4, "stem_gpqa_score": 67.7, "osworld_score": 31, "human_elo": 1222, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.14, "pricing_output_per_m": 0.45, "cache_read_per_m": 0.014, "reality_gap": 7.1, "category": "Coding & Agentic", "key_moat": "Evaluated constituent #103 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://salesforce.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 104, "model_id": "modl-upstage-v2-104", "name": "Upstage Model-104 (Enterprise & Math)", "developer": "Upstage", "license_type": "open_weights", "composite_iq_score": 88.4, "reasoning_score": 88.1, "agent_code_score": 53.2, "stem_gpqa_score": 67.6, "osworld_score": 30.9, "human_elo": 1221, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.16, "pricing_output_per_m": 0.51, "cache_read_per_m": 0.016, "reality_gap": 8.2, "category": "Enterprise & Math", "key_moat": "Evaluated constituent #104 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://upstage.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 105, "model_id": "modl-deci-ai-v3-105", "name": "Deci AI Model-105 (Multimodal & Vision)", "developer": "Deci AI", "license_type": "proprietary", "composite_iq_score": 88.3, "reasoning_score": 88, "agent_code_score": 53.1, "stem_gpqa_score": 67.5, "osworld_score": 30.8, "human_elo": 1220, "multimodal_score": 91, "context_window": "64k", "pricing_input_per_m": 0.7, "pricing_output_per_m": 2.24, "cache_read_per_m": 0.07, "reality_gap": 9.3, "category": "Multimodal & Vision", "key_moat": "Evaluated constituent #105 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://deciai.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 106, "model_id": "modl-mistral-ai-v4-106", "name": "Mistral AI Model-106 (Efficiency & Edge)", "developer": "Mistral AI", "license_type": "open_weights", "composite_iq_score": 88.3, "reasoning_score": 88, "agent_code_score": 53, "stem_gpqa_score": 67.3, "osworld_score": 30.7, "human_elo": 1220, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.1, "pricing_output_per_m": 0.32, "cache_read_per_m": 0.01, "reality_gap": 10.4, "category": "Efficiency & Edge", "key_moat": "Evaluated constituent #106 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://mistralai.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 107, "model_id": "modl-qwen-team-v5-107", "name": "Qwen Team Model-107 (Open-Weights SOTA)", "developer": "Qwen Team", "license_type": "open_weights", "composite_iq_score": 88.2, "reasoning_score": 87.9, "agent_code_score": 52.8, "stem_gpqa_score": 67.2, "osworld_score": 30.6, "human_elo": 1218, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.12, "pricing_output_per_m": 0.38, "cache_read_per_m": 0.012, "reality_gap": 11.5, "category": "Open-Weights SOTA", "key_moat": "Evaluated constituent #107 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://qwenteam.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 108, "model_id": "modl-deepseek-v6-108", "name": "DeepSeek Model-108 (Coding & Agentic)", "developer": "DeepSeek", "license_type": "proprietary", "composite_iq_score": 88.1, "reasoning_score": 87.8, "agent_code_score": 52.7, "stem_gpqa_score": 67.1, "osworld_score": 30.4, "human_elo": 1217, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.69, "pricing_output_per_m": 2.21, "cache_read_per_m": 0.069, "reality_gap": 6, "category": "Coding & Agentic", "key_moat": "Evaluated constituent #108 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://deepseek.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 109, "model_id": "modl-meta-v7-109", "name": "Meta Model-109 (Enterprise & Math)", "developer": "Meta", "license_type": "open_weights", "composite_iq_score": 88, "reasoning_score": 87.7, "agent_code_score": 52.5, "stem_gpqa_score": 66.9, "osworld_score": 30.3, "human_elo": 1216, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.16, "pricing_output_per_m": 0.51, "cache_read_per_m": 0.016, "reality_gap": 7.1, "category": "Enterprise & Math", "key_moat": "Evaluated constituent #109 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://meta.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 110, "model_id": "modl-google-deepmind-v8-110", "name": "Google DeepMind Model-110 (Multimodal & Vision)", "developer": "Google DeepMind", "license_type": "open_weights", "composite_iq_score": 87.9, "reasoning_score": 87.6, "agent_code_score": 52.4, "stem_gpqa_score": 66.8, "osworld_score": 30.2, "human_elo": 1215, "multimodal_score": 91, "context_window": "128k", "pricing_input_per_m": 0.08, "pricing_output_per_m": 0.26, "cache_read_per_m": 0.008, "reality_gap": 8.2, "category": "Multimodal & Vision", "key_moat": "Evaluated constituent #110 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://googledeepmind.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 111, "model_id": "modl-anthropic-v9-111", "name": "Anthropic Model-111 (Efficiency & Edge)", "developer": "Anthropic", "license_type": "proprietary", "composite_iq_score": 87.8, "reasoning_score": 87.5, "agent_code_score": 52.3, "stem_gpqa_score": 66.7, "osworld_score": 30.1, "human_elo": 1214, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.68, "pricing_output_per_m": 2.18, "cache_read_per_m": 0.068, "reality_gap": 9.3, "category": "Efficiency & Edge", "key_moat": "Evaluated constituent #111 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://anthropic.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 112, "model_id": "modl-openai-v10-112", "name": "OpenAI Model-112 (Open-Weights SOTA)", "developer": "OpenAI", "license_type": "open_weights", "composite_iq_score": 87.7, "reasoning_score": 87.4, "agent_code_score": 52.1, "stem_gpqa_score": 66.5, "osworld_score": 30, "human_elo": 1212, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.12, "pricing_output_per_m": 0.38, "cache_read_per_m": 0.012, "reality_gap": 10.4, "category": "Open-Weights SOTA", "key_moat": "Evaluated constituent #112 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://openai.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 113, "model_id": "modl-cohere-v11-113", "name": "Cohere Model-113 (Coding & Agentic)", "developer": "Cohere", "license_type": "open_weights", "composite_iq_score": 87.6, "reasoning_score": 87.3, "agent_code_score": 52, "stem_gpqa_score": 66.4, "osworld_score": 29.8, "human_elo": 1211, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.14, "pricing_output_per_m": 0.45, "cache_read_per_m": 0.014, "reality_gap": 11.5, "category": "Coding & Agentic", "key_moat": "Evaluated constituent #113 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://cohere.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 114, "model_id": "modl-ai21-labs-v12-114", "name": "AI21 Labs Model-114 (Enterprise & Math)", "developer": "AI21 Labs", "license_type": "proprietary", "composite_iq_score": 87.5, "reasoning_score": 87.2, "agent_code_score": 51.8, "stem_gpqa_score": 66.3, "osworld_score": 29.7, "human_elo": 1210, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.67, "pricing_output_per_m": 2.14, "cache_read_per_m": 0.067, "reality_gap": 6, "category": "Enterprise & Math", "key_moat": "Evaluated constituent #114 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://ai21labs.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 115, "model_id": "modl-tii-falcon-v13-115", "name": "TII Falcon Model-115 (Multimodal & Vision)", "developer": "TII Falcon", "license_type": "open_weights", "composite_iq_score": 87.5, "reasoning_score": 87.2, "agent_code_score": 51.7, "stem_gpqa_score": 66.2, "osworld_score": 29.6, "human_elo": 1210, "multimodal_score": 91, "context_window": "64k", "pricing_input_per_m": 0.08, "pricing_output_per_m": 0.26, "cache_read_per_m": 0.008, "reality_gap": 7.1, "category": "Multimodal & Vision", "key_moat": "Evaluated constituent #115 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://tiifalcon.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 116, "model_id": "modl-aleph-alpha-v14-116", "name": "Aleph Alpha Model-116 (Efficiency & Edge)", "developer": "Aleph Alpha", "license_type": "open_weights", "composite_iq_score": 87.4, "reasoning_score": 87.1, "agent_code_score": 51.6, "stem_gpqa_score": 66, "osworld_score": 29.5, "human_elo": 1209, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.1, "pricing_output_per_m": 0.32, "cache_read_per_m": 0.01, "reality_gap": 8.2, "category": "Efficiency & Edge", "key_moat": "Evaluated constituent #116 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://alephalpha.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 117, "model_id": "modl-stability-ai-v15-117", "name": "Stability AI Model-117 (Open-Weights SOTA)", "developer": "Stability AI", "license_type": "proprietary", "composite_iq_score": 87.3, "reasoning_score": 87, "agent_code_score": 51.4, "stem_gpqa_score": 65.9, "osworld_score": 29.4, "human_elo": 1208, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.66, "pricing_output_per_m": 2.11, "cache_read_per_m": 0.066, "reality_gap": 9.3, "category": "Open-Weights SOTA", "key_moat": "Evaluated constituent #117 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://stabilityai.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 118, "model_id": "modl-salesforce-v16-118", "name": "Salesforce Model-118 (Coding & Agentic)", "developer": "Salesforce", "license_type": "open_weights", "composite_iq_score": 87.2, "reasoning_score": 86.9, "agent_code_score": 51.3, "stem_gpqa_score": 65.8, "osworld_score": 29.2, "human_elo": 1206, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.14, "pricing_output_per_m": 0.45, "cache_read_per_m": 0.014, "reality_gap": 10.4, "category": "Coding & Agentic", "key_moat": "Evaluated constituent #118 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://salesforce.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 119, "model_id": "modl-upstage-v17-119", "name": "Upstage Model-119 (Enterprise & Math)", "developer": "Upstage", "license_type": "open_weights", "composite_iq_score": 87.1, "reasoning_score": 86.8, "agent_code_score": 51.1, "stem_gpqa_score": 65.6, "osworld_score": 29.1, "human_elo": 1205, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.16, "pricing_output_per_m": 0.51, "cache_read_per_m": 0.016, "reality_gap": 11.5, "category": "Enterprise & Math", "key_moat": "Evaluated constituent #119 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://upstage.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 120, "model_id": "modl-deci-ai-v18-120", "name": "Deci AI Model-120 (Multimodal & Vision)", "developer": "Deci AI", "license_type": "proprietary", "composite_iq_score": 87, "reasoning_score": 86.7, "agent_code_score": 51, "stem_gpqa_score": 65.5, "osworld_score": 29, "human_elo": 1204, "multimodal_score": 91, "context_window": "128k", "pricing_input_per_m": 0.65, "pricing_output_per_m": 2.08, "cache_read_per_m": 0.065, "reality_gap": 6, "category": "Multimodal & Vision", "key_moat": "Evaluated constituent #120 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://deciai.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 121, "model_id": "modl-mistral-ai-v19-121", "name": "Mistral AI Model-121 (Efficiency & Edge)", "developer": "Mistral AI", "license_type": "open_weights", "composite_iq_score": 86.9, "reasoning_score": 86.6, "agent_code_score": 50.9, "stem_gpqa_score": 65.4, "osworld_score": 28.9, "human_elo": 1203, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.1, "pricing_output_per_m": 0.32, "cache_read_per_m": 0.01, "reality_gap": 7.1, "category": "Efficiency & Edge", "key_moat": "Evaluated constituent #121 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://mistralai.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 122, "model_id": "modl-qwen-team-v20-122", "name": "Qwen Team Model-122 (Open-Weights SOTA)", "developer": "Qwen Team", "license_type": "open_weights", "composite_iq_score": 86.8, "reasoning_score": 86.5, "agent_code_score": 50.7, "stem_gpqa_score": 65.2, "osworld_score": 28.8, "human_elo": 1202, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.12, "pricing_output_per_m": 0.38, "cache_read_per_m": 0.012, "reality_gap": 8.2, "category": "Open-Weights SOTA", "key_moat": "Evaluated constituent #122 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://qwenteam.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 123, "model_id": "modl-deepseek-v21-123", "name": "DeepSeek Model-123 (Coding & Agentic)", "developer": "DeepSeek", "license_type": "proprietary", "composite_iq_score": 86.7, "reasoning_score": 86.4, "agent_code_score": 50.6, "stem_gpqa_score": 65.1, "osworld_score": 28.6, "human_elo": 1200, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.64, "pricing_output_per_m": 2.05, "cache_read_per_m": 0.064, "reality_gap": 9.3, "category": "Coding & Agentic", "key_moat": "Evaluated constituent #123 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://deepseek.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 124, "model_id": "modl-meta-v22-124", "name": "Meta Model-124 (Enterprise & Math)", "developer": "Meta", "license_type": "open_weights", "composite_iq_score": 86.6, "reasoning_score": 86.3, "agent_code_score": 50.4, "stem_gpqa_score": 65, "osworld_score": 28.5, "human_elo": 1199, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.16, "pricing_output_per_m": 0.51, "cache_read_per_m": 0.016, "reality_gap": 10.4, "category": "Enterprise & Math", "key_moat": "Evaluated constituent #124 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://meta.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 125, "model_id": "modl-google-deepmind-v23-125", "name": "Google DeepMind Model-125 (Multimodal & Vision)", "developer": "Google DeepMind", "license_type": "open_weights", "composite_iq_score": 86.5, "reasoning_score": 86.2, "agent_code_score": 50.3, "stem_gpqa_score": 64.8, "osworld_score": 28.4, "human_elo": 1198, "multimodal_score": 91, "context_window": "64k", "pricing_input_per_m": 0.08, "pricing_output_per_m": 0.26, "cache_read_per_m": 0.008, "reality_gap": 11.5, "category": "Multimodal & Vision", "key_moat": "Evaluated constituent #125 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://googledeepmind.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 126, "model_id": "modl-anthropic-v24-126", "name": "Anthropic Model-126 (Efficiency & Edge)", "developer": "Anthropic", "license_type": "proprietary", "composite_iq_score": 86.5, "reasoning_score": 86.2, "agent_code_score": 50.2, "stem_gpqa_score": 64.7, "osworld_score": 28.3, "human_elo": 1198, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.63, "pricing_output_per_m": 2.02, "cache_read_per_m": 0.063, "reality_gap": 6, "category": "Efficiency & Edge", "key_moat": "Evaluated constituent #126 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://anthropic.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 127, "model_id": "modl-openai-v25-127", "name": "OpenAI Model-127 (Open-Weights SOTA)", "developer": "OpenAI", "license_type": "open_weights", "composite_iq_score": 86.4, "reasoning_score": 86.1, "agent_code_score": 50, "stem_gpqa_score": 64.6, "osworld_score": 28.2, "human_elo": 1197, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.12, "pricing_output_per_m": 0.38, "cache_read_per_m": 0.012, "reality_gap": 7.1, "category": "Open-Weights SOTA", "key_moat": "Evaluated constituent #127 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://openai.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 128, "model_id": "modl-cohere-v26-128", "name": "Cohere Model-128 (Coding & Agentic)", "developer": "Cohere", "license_type": "open_weights", "composite_iq_score": 86.3, "reasoning_score": 86, "agent_code_score": 49.9, "stem_gpqa_score": 64.5, "osworld_score": 28, "human_elo": 1196, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.14, "pricing_output_per_m": 0.45, "cache_read_per_m": 0.014, "reality_gap": 8.2, "category": "Coding & Agentic", "key_moat": "Evaluated constituent #128 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://cohere.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 129, "model_id": "modl-ai21-labs-v27-129", "name": "AI21 Labs Model-129 (Enterprise & Math)", "developer": "AI21 Labs", "license_type": "proprietary", "composite_iq_score": 86.2, "reasoning_score": 85.9, "agent_code_score": 49.7, "stem_gpqa_score": 64.3, "osworld_score": 27.9, "human_elo": 1194, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.62, "pricing_output_per_m": 1.98, "cache_read_per_m": 0.062, "reality_gap": 9.3, "category": "Enterprise & Math", "key_moat": "Evaluated constituent #129 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://ai21labs.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 130, "model_id": "modl-tii-falcon-v28-130", "name": "TII Falcon Model-130 (Multimodal & Vision)", "developer": "TII Falcon", "license_type": "open_weights", "composite_iq_score": 86.1, "reasoning_score": 85.8, "agent_code_score": 49.6, "stem_gpqa_score": 64.2, "osworld_score": 27.8, "human_elo": 1193, "multimodal_score": 91, "context_window": "128k", "pricing_input_per_m": 0.08, "pricing_output_per_m": 0.26, "cache_read_per_m": 0.008, "reality_gap": 10.4, "category": "Multimodal & Vision", "key_moat": "Evaluated constituent #130 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://tiifalcon.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 131, "model_id": "modl-aleph-alpha-v29-131", "name": "Aleph Alpha Model-131 (Efficiency & Edge)", "developer": "Aleph Alpha", "license_type": "open_weights", "composite_iq_score": 86, "reasoning_score": 85.7, "agent_code_score": 49.5, "stem_gpqa_score": 64.1, "osworld_score": 27.7, "human_elo": 1192, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.1, "pricing_output_per_m": 0.32, "cache_read_per_m": 0.01, "reality_gap": 11.5, "category": "Efficiency & Edge", "key_moat": "Evaluated constituent #131 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://alephalpha.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 132, "model_id": "modl-stability-ai-v30-132", "name": "Stability AI Model-132 (Open-Weights SOTA)", "developer": "Stability AI", "license_type": "proprietary", "composite_iq_score": 85.9, "reasoning_score": 85.6, "agent_code_score": 49.3, "stem_gpqa_score": 63.9, "osworld_score": 27.6, "human_elo": 1191, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.61, "pricing_output_per_m": 1.95, "cache_read_per_m": 0.061, "reality_gap": 6, "category": "Open-Weights SOTA", "key_moat": "Evaluated constituent #132 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://stabilityai.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 133, "model_id": "modl-salesforce-v31-133", "name": "Salesforce Model-133 (Coding & Agentic)", "developer": "Salesforce", "license_type": "open_weights", "composite_iq_score": 85.8, "reasoning_score": 85.5, "agent_code_score": 49.2, "stem_gpqa_score": 63.8, "osworld_score": 27.4, "human_elo": 1190, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.14, "pricing_output_per_m": 0.45, "cache_read_per_m": 0.014, "reality_gap": 7.1, "category": "Coding & Agentic", "key_moat": "Evaluated constituent #133 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://salesforce.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 134, "model_id": "modl-upstage-v32-134", "name": "Upstage Model-134 (Enterprise & Math)", "developer": "Upstage", "license_type": "open_weights", "composite_iq_score": 85.7, "reasoning_score": 85.4, "agent_code_score": 49, "stem_gpqa_score": 63.7, "osworld_score": 27.3, "human_elo": 1188, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.16, "pricing_output_per_m": 0.51, "cache_read_per_m": 0.016, "reality_gap": 8.2, "category": "Enterprise & Math", "key_moat": "Evaluated constituent #134 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://upstage.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 135, "model_id": "modl-deci-ai-v33-135", "name": "Deci AI Model-135 (Multimodal & Vision)", "developer": "Deci AI", "license_type": "proprietary", "composite_iq_score": 85.7, "reasoning_score": 85.4, "agent_code_score": 48.9, "stem_gpqa_score": 63.5, "osworld_score": 27.2, "human_elo": 1188, "multimodal_score": 91, "context_window": "64k", "pricing_input_per_m": 0.6, "pricing_output_per_m": 1.92, "cache_read_per_m": 0.06, "reality_gap": 9.3, "category": "Multimodal & Vision", "key_moat": "Evaluated constituent #135 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://deciai.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 136, "model_id": "modl-mistral-ai-v34-136", "name": "Mistral AI Model-136 (Efficiency & Edge)", "developer": "Mistral AI", "license_type": "open_weights", "composite_iq_score": 85.6, "reasoning_score": 85.3, "agent_code_score": 48.8, "stem_gpqa_score": 63.4, "osworld_score": 27.1, "human_elo": 1187, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.1, "pricing_output_per_m": 0.32, "cache_read_per_m": 0.01, "reality_gap": 10.4, "category": "Efficiency & Edge", "key_moat": "Evaluated constituent #136 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://mistralai.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 137, "model_id": "modl-qwen-team-v35-137", "name": "Qwen Team Model-137 (Open-Weights SOTA)", "developer": "Qwen Team", "license_type": "open_weights", "composite_iq_score": 85.5, "reasoning_score": 85.2, "agent_code_score": 48.6, "stem_gpqa_score": 63.3, "osworld_score": 27, "human_elo": 1186, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.12, "pricing_output_per_m": 0.38, "cache_read_per_m": 0.012, "reality_gap": 11.5, "category": "Open-Weights SOTA", "key_moat": "Evaluated constituent #137 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://qwenteam.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 138, "model_id": "modl-deepseek-v36-138", "name": "DeepSeek Model-138 (Coding & Agentic)", "developer": "DeepSeek", "license_type": "proprietary", "composite_iq_score": 85.4, "reasoning_score": 85.1, "agent_code_score": 48.5, "stem_gpqa_score": 63.2, "osworld_score": 26.8, "human_elo": 1185, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.6, "pricing_output_per_m": 1.92, "cache_read_per_m": 0.06, "reality_gap": 6, "category": "Coding & Agentic", "key_moat": "Evaluated constituent #138 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://deepseek.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 139, "model_id": "modl-meta-v37-139", "name": "Meta Model-139 (Enterprise & Math)", "developer": "Meta", "license_type": "open_weights", "composite_iq_score": 85.3, "reasoning_score": 85, "agent_code_score": 48.3, "stem_gpqa_score": 63, "osworld_score": 26.7, "human_elo": 1184, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.16, "pricing_output_per_m": 0.51, "cache_read_per_m": 0.016, "reality_gap": 7.1, "category": "Enterprise & Math", "key_moat": "Evaluated constituent #139 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://meta.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 140, "model_id": "modl-google-deepmind-v38-140", "name": "Google DeepMind Model-140 (Multimodal & Vision)", "developer": "Google DeepMind", "license_type": "open_weights", "composite_iq_score": 85.2, "reasoning_score": 84.9, "agent_code_score": 48.2, "stem_gpqa_score": 62.9, "osworld_score": 26.6, "human_elo": 1182, "multimodal_score": 91, "context_window": "128k", "pricing_input_per_m": 0.08, "pricing_output_per_m": 0.26, "cache_read_per_m": 0.008, "reality_gap": 8.2, "category": "Multimodal & Vision", "key_moat": "Evaluated constituent #140 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://googledeepmind.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 141, "model_id": "modl-anthropic-v39-141", "name": "Anthropic Model-141 (Efficiency & Edge)", "developer": "Anthropic", "license_type": "proprietary", "composite_iq_score": 85.1, "reasoning_score": 84.8, "agent_code_score": 48.1, "stem_gpqa_score": 62.8, "osworld_score": 26.5, "human_elo": 1181, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.59, "pricing_output_per_m": 1.89, "cache_read_per_m": 0.059, "reality_gap": 9.3, "category": "Efficiency & Edge", "key_moat": "Evaluated constituent #141 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://anthropic.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 142, "model_id": "modl-openai-v40-142", "name": "OpenAI Model-142 (Open-Weights SOTA)", "developer": "OpenAI", "license_type": "open_weights", "composite_iq_score": 85, "reasoning_score": 84.7, "agent_code_score": 47.9, "stem_gpqa_score": 62.6, "osworld_score": 26.4, "human_elo": 1180, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.12, "pricing_output_per_m": 0.38, "cache_read_per_m": 0.012, "reality_gap": 10.4, "category": "Open-Weights SOTA", "key_moat": "Evaluated constituent #142 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://openai.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 143, "model_id": "modl-cohere-v41-143", "name": "Cohere Model-143 (Coding & Agentic)", "developer": "Cohere", "license_type": "open_weights", "composite_iq_score": 84.9, "reasoning_score": 84.6, "agent_code_score": 47.8, "stem_gpqa_score": 62.5, "osworld_score": 26.2, "human_elo": 1179, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.14, "pricing_output_per_m": 0.45, "cache_read_per_m": 0.014, "reality_gap": 11.5, "category": "Coding & Agentic", "key_moat": "Evaluated constituent #143 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://cohere.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 144, "model_id": "modl-ai21-labs-v42-144", "name": "AI21 Labs Model-144 (Enterprise & Math)", "developer": "AI21 Labs", "license_type": "proprietary", "composite_iq_score": 84.8, "reasoning_score": 84.5, "agent_code_score": 47.6, "stem_gpqa_score": 62.4, "osworld_score": 26.1, "human_elo": 1178, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.58, "pricing_output_per_m": 1.86, "cache_read_per_m": 0.058, "reality_gap": 6, "category": "Enterprise & Math", "key_moat": "Evaluated constituent #144 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://ai21labs.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 145, "model_id": "modl-tii-falcon-v43-145", "name": "TII Falcon Model-145 (Multimodal & Vision)", "developer": "TII Falcon", "license_type": "open_weights", "composite_iq_score": 84.8, "reasoning_score": 84.5, "agent_code_score": 47.5, "stem_gpqa_score": 62.3, "osworld_score": 26, "human_elo": 1178, "multimodal_score": 91, "context_window": "64k", "pricing_input_per_m": 0.08, "pricing_output_per_m": 0.26, "cache_read_per_m": 0.008, "reality_gap": 7.1, "category": "Multimodal & Vision", "key_moat": "Evaluated constituent #145 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://tiifalcon.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 146, "model_id": "modl-aleph-alpha-v44-146", "name": "Aleph Alpha Model-146 (Efficiency & Edge)", "developer": "Aleph Alpha", "license_type": "open_weights", "composite_iq_score": 84.7, "reasoning_score": 84.4, "agent_code_score": 47.4, "stem_gpqa_score": 62.1, "osworld_score": 25.9, "human_elo": 1176, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.1, "pricing_output_per_m": 0.32, "cache_read_per_m": 0.01, "reality_gap": 8.2, "category": "Efficiency & Edge", "key_moat": "Evaluated constituent #146 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://alephalpha.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 147, "model_id": "modl-stability-ai-v45-147", "name": "Stability AI Model-147 (Open-Weights SOTA)", "developer": "Stability AI", "license_type": "proprietary", "composite_iq_score": 84.6, "reasoning_score": 84.3, "agent_code_score": 47.2, "stem_gpqa_score": 62, "osworld_score": 25.8, "human_elo": 1175, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.57, "pricing_output_per_m": 1.82, "cache_read_per_m": 0.057, "reality_gap": 9.3, "category": "Open-Weights SOTA", "key_moat": "Evaluated constituent #147 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://stabilityai.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 148, "model_id": "modl-salesforce-v46-148", "name": "Salesforce Model-148 (Coding & Agentic)", "developer": "Salesforce", "license_type": "open_weights", "composite_iq_score": 84.5, "reasoning_score": 84.2, "agent_code_score": 47.1, "stem_gpqa_score": 61.9, "osworld_score": 25.6, "human_elo": 1174, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.14, "pricing_output_per_m": 0.45, "cache_read_per_m": 0.014, "reality_gap": 10.4, "category": "Coding & Agentic", "key_moat": "Evaluated constituent #148 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://salesforce.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 149, "model_id": "modl-upstage-v47-149", "name": "Upstage Model-149 (Enterprise & Math)", "developer": "Upstage", "license_type": "open_weights", "composite_iq_score": 84.4, "reasoning_score": 84.1, "agent_code_score": 46.9, "stem_gpqa_score": 61.7, "osworld_score": 25.5, "human_elo": 1173, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.16, "pricing_output_per_m": 0.51, "cache_read_per_m": 0.016, "reality_gap": 11.5, "category": "Enterprise & Math", "key_moat": "Evaluated constituent #149 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://upstage.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 150, "model_id": "modl-deci-ai-v48-150", "name": "Deci AI Model-150 (Multimodal & Vision)", "developer": "Deci AI", "license_type": "proprietary", "composite_iq_score": 84.3, "reasoning_score": 84, "agent_code_score": 46.8, "stem_gpqa_score": 61.6, "osworld_score": 25.4, "human_elo": 1172, "multimodal_score": 91, "context_window": "128k", "pricing_input_per_m": 0.56, "pricing_output_per_m": 1.79, "cache_read_per_m": 0.056, "reality_gap": 6, "category": "Multimodal & Vision", "key_moat": "Evaluated constituent #150 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://deciai.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 151, "model_id": "modl-mistral-ai-v49-151", "name": "Mistral AI Model-151 (Efficiency & Edge)", "developer": "Mistral AI", "license_type": "open_weights", "composite_iq_score": 84.2, "reasoning_score": 83.9, "agent_code_score": 46.7, "stem_gpqa_score": 61.5, "osworld_score": 25.3, "human_elo": 1170, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.1, "pricing_output_per_m": 0.32, "cache_read_per_m": 0.01, "reality_gap": 7.1, "category": "Efficiency & Edge", "key_moat": "Evaluated constituent #151 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://mistralai.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 152, "model_id": "modl-qwen-team-v50-152", "name": "Qwen Team Model-152 (Open-Weights SOTA)", "developer": "Qwen Team", "license_type": "open_weights", "composite_iq_score": 84.1, "reasoning_score": 83.8, "agent_code_score": 46.5, "stem_gpqa_score": 61.3, "osworld_score": 25.2, "human_elo": 1169, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.12, "pricing_output_per_m": 0.38, "cache_read_per_m": 0.012, "reality_gap": 8.2, "category": "Open-Weights SOTA", "key_moat": "Evaluated constituent #152 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://qwenteam.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 153, "model_id": "modl-deepseek-v51-153", "name": "DeepSeek Model-153 (Coding & Agentic)", "developer": "DeepSeek", "license_type": "proprietary", "composite_iq_score": 84, "reasoning_score": 83.7, "agent_code_score": 46.4, "stem_gpqa_score": 61.2, "osworld_score": 25, "human_elo": 1168, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.55, "pricing_output_per_m": 1.76, "cache_read_per_m": 0.055, "reality_gap": 9.3, "category": "Coding & Agentic", "key_moat": "Evaluated constituent #153 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://deepseek.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 154, "model_id": "modl-meta-v52-154", "name": "Meta Model-154 (Enterprise & Math)", "developer": "Meta", "license_type": "open_weights", "composite_iq_score": 83.9, "reasoning_score": 83.6, "agent_code_score": 46.2, "stem_gpqa_score": 61.1, "osworld_score": 24.9, "human_elo": 1167, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.16, "pricing_output_per_m": 0.51, "cache_read_per_m": 0.016, "reality_gap": 10.4, "category": "Enterprise & Math", "key_moat": "Evaluated constituent #154 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://meta.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 155, "model_id": "modl-google-deepmind-v53-155", "name": "Google DeepMind Model-155 (Multimodal & Vision)", "developer": "Google DeepMind", "license_type": "open_weights", "composite_iq_score": 83.8, "reasoning_score": 83.5, "agent_code_score": 46.1, "stem_gpqa_score": 61, "osworld_score": 24.8, "human_elo": 1166, "multimodal_score": 91, "context_window": "64k", "pricing_input_per_m": 0.08, "pricing_output_per_m": 0.26, "cache_read_per_m": 0.008, "reality_gap": 11.5, "category": "Multimodal & Vision", "key_moat": "Evaluated constituent #155 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://googledeepmind.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 156, "model_id": "modl-anthropic-v54-156", "name": "Anthropic Model-156 (Efficiency & Edge)", "developer": "Anthropic", "license_type": "proprietary", "composite_iq_score": 83.8, "reasoning_score": 83.5, "agent_code_score": 46, "stem_gpqa_score": 60.8, "osworld_score": 24.7, "human_elo": 1166, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.54, "pricing_output_per_m": 1.73, "cache_read_per_m": 0.054, "reality_gap": 6, "category": "Efficiency & Edge", "key_moat": "Evaluated constituent #156 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://anthropic.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 157, "model_id": "modl-openai-v55-157", "name": "OpenAI Model-157 (Open-Weights SOTA)", "developer": "OpenAI", "license_type": "open_weights", "composite_iq_score": 83.7, "reasoning_score": 83.4, "agent_code_score": 45.8, "stem_gpqa_score": 60.7, "osworld_score": 24.6, "human_elo": 1164, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.12, "pricing_output_per_m": 0.38, "cache_read_per_m": 0.012, "reality_gap": 7.1, "category": "Open-Weights SOTA", "key_moat": "Evaluated constituent #157 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://openai.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 158, "model_id": "modl-cohere-v56-158", "name": "Cohere Model-158 (Coding & Agentic)", "developer": "Cohere", "license_type": "open_weights", "composite_iq_score": 83.6, "reasoning_score": 83.3, "agent_code_score": 45.7, "stem_gpqa_score": 60.6, "osworld_score": 24.4, "human_elo": 1163, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.14, "pricing_output_per_m": 0.45, "cache_read_per_m": 0.014, "reality_gap": 8.2, "category": "Coding & Agentic", "key_moat": "Evaluated constituent #158 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://cohere.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 159, "model_id": "modl-ai21-labs-v57-159", "name": "AI21 Labs Model-159 (Enterprise & Math)", "developer": "AI21 Labs", "license_type": "proprietary", "composite_iq_score": 83.5, "reasoning_score": 83.2, "agent_code_score": 45.5, "stem_gpqa_score": 60.4, "osworld_score": 24.3, "human_elo": 1162, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.53, "pricing_output_per_m": 1.7, "cache_read_per_m": 0.053, "reality_gap": 9.3, "category": "Enterprise & Math", "key_moat": "Evaluated constituent #159 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://ai21labs.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 160, "model_id": "modl-tii-falcon-v58-160", "name": "TII Falcon Model-160 (Multimodal & Vision)", "developer": "TII Falcon", "license_type": "open_weights", "composite_iq_score": 83.4, "reasoning_score": 83.1, "agent_code_score": 45.4, "stem_gpqa_score": 60.3, "osworld_score": 24.2, "human_elo": 1161, "multimodal_score": 91, "context_window": "128k", "pricing_input_per_m": 0.08, "pricing_output_per_m": 0.26, "cache_read_per_m": 0.008, "reality_gap": 10.4, "category": "Multimodal & Vision", "key_moat": "Evaluated constituent #160 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://tiifalcon.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 161, "model_id": "modl-aleph-alpha-v59-161", "name": "Aleph Alpha Model-161 (Efficiency & Edge)", "developer": "Aleph Alpha", "license_type": "open_weights", "composite_iq_score": 83.3, "reasoning_score": 83, "agent_code_score": 45.3, "stem_gpqa_score": 60.2, "osworld_score": 24.1, "human_elo": 1160, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.1, "pricing_output_per_m": 0.32, "cache_read_per_m": 0.01, "reality_gap": 11.5, "category": "Efficiency & Edge", "key_moat": "Evaluated constituent #161 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://alephalpha.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 162, "model_id": "modl-stability-ai-v60-162", "name": "Stability AI Model-162 (Open-Weights SOTA)", "developer": "Stability AI", "license_type": "proprietary", "composite_iq_score": 83.2, "reasoning_score": 82.9, "agent_code_score": 45.1, "stem_gpqa_score": 60, "osworld_score": 24, "human_elo": 1158, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.52, "pricing_output_per_m": 1.66, "cache_read_per_m": 0.052, "reality_gap": 6, "category": "Open-Weights SOTA", "key_moat": "Evaluated constituent #162 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://stabilityai.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 163, "model_id": "modl-salesforce-v61-163", "name": "Salesforce Model-163 (Coding & Agentic)", "developer": "Salesforce", "license_type": "open_weights", "composite_iq_score": 83.1, "reasoning_score": 82.8, "agent_code_score": 45, "stem_gpqa_score": 59.9, "osworld_score": 23.8, "human_elo": 1157, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.14, "pricing_output_per_m": 0.45, "cache_read_per_m": 0.014, "reality_gap": 7.1, "category": "Coding & Agentic", "key_moat": "Evaluated constituent #163 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://salesforce.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 164, "model_id": "modl-upstage-v62-164", "name": "Upstage Model-164 (Enterprise & Math)", "developer": "Upstage", "license_type": "open_weights", "composite_iq_score": 83, "reasoning_score": 82.7, "agent_code_score": 44.8, "stem_gpqa_score": 59.8, "osworld_score": 23.7, "human_elo": 1156, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.16, "pricing_output_per_m": 0.51, "cache_read_per_m": 0.016, "reality_gap": 8.2, "category": "Enterprise & Math", "key_moat": "Evaluated constituent #164 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://upstage.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 165, "model_id": "modl-deci-ai-v63-165", "name": "Deci AI Model-165 (Multimodal & Vision)", "developer": "Deci AI", "license_type": "proprietary", "composite_iq_score": 83, "reasoning_score": 82.7, "agent_code_score": 44.7, "stem_gpqa_score": 59.6, "osworld_score": 23.6, "human_elo": 1156, "multimodal_score": 91, "context_window": "64k", "pricing_input_per_m": 0.52, "pricing_output_per_m": 1.66, "cache_read_per_m": 0.052, "reality_gap": 9.3, "category": "Multimodal & Vision", "key_moat": "Evaluated constituent #165 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://deciai.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 166, "model_id": "modl-mistral-ai-v64-166", "name": "Mistral AI Model-166 (Efficiency & Edge)", "developer": "Mistral AI", "license_type": "open_weights", "composite_iq_score": 82.9, "reasoning_score": 82.6, "agent_code_score": 44.6, "stem_gpqa_score": 59.5, "osworld_score": 23.5, "human_elo": 1155, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.1, "pricing_output_per_m": 0.32, "cache_read_per_m": 0.01, "reality_gap": 10.4, "category": "Efficiency & Edge", "key_moat": "Evaluated constituent #166 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://mistralai.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 167, "model_id": "modl-qwen-team-v65-167", "name": "Qwen Team Model-167 (Open-Weights SOTA)", "developer": "Qwen Team", "license_type": "open_weights", "composite_iq_score": 82.8, "reasoning_score": 82.5, "agent_code_score": 44.4, "stem_gpqa_score": 59.4, "osworld_score": 23.4, "human_elo": 1154, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.12, "pricing_output_per_m": 0.38, "cache_read_per_m": 0.012, "reality_gap": 11.5, "category": "Open-Weights SOTA", "key_moat": "Evaluated constituent #167 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://qwenteam.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 168, "model_id": "modl-deepseek-v66-168", "name": "DeepSeek Model-168 (Coding & Agentic)", "developer": "DeepSeek", "license_type": "proprietary", "composite_iq_score": 82.7, "reasoning_score": 82.4, "agent_code_score": 44.3, "stem_gpqa_score": 59.3, "osworld_score": 23.2, "human_elo": 1152, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.51, "pricing_output_per_m": 1.63, "cache_read_per_m": 0.051, "reality_gap": 6, "category": "Coding & Agentic", "key_moat": "Evaluated constituent #168 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://deepseek.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 169, "model_id": "modl-meta-v67-169", "name": "Meta Model-169 (Enterprise & Math)", "developer": "Meta", "license_type": "open_weights", "composite_iq_score": 82.6, "reasoning_score": 82.3, "agent_code_score": 44.1, "stem_gpqa_score": 59.1, "osworld_score": 23.1, "human_elo": 1151, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.16, "pricing_output_per_m": 0.51, "cache_read_per_m": 0.016, "reality_gap": 7.1, "category": "Enterprise & Math", "key_moat": "Evaluated constituent #169 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://meta.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 170, "model_id": "modl-google-deepmind-v68-170", "name": "Google DeepMind Model-170 (Multimodal & Vision)", "developer": "Google DeepMind", "license_type": "open_weights", "composite_iq_score": 82.5, "reasoning_score": 82.2, "agent_code_score": 44, "stem_gpqa_score": 59, "osworld_score": 23, "human_elo": 1150, "multimodal_score": 91, "context_window": "128k", "pricing_input_per_m": 0.08, "pricing_output_per_m": 0.26, "cache_read_per_m": 0.008, "reality_gap": 8.2, "category": "Multimodal & Vision", "key_moat": "Evaluated constituent #170 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://googledeepmind.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 171, "model_id": "modl-anthropic-v69-171", "name": "Anthropic Model-171 (Efficiency & Edge)", "developer": "Anthropic", "license_type": "proprietary", "composite_iq_score": 82.4, "reasoning_score": 82.1, "agent_code_score": 43.9, "stem_gpqa_score": 58.9, "osworld_score": 22.9, "human_elo": 1149, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.5, "pricing_output_per_m": 1.6, "cache_read_per_m": 0.05, "reality_gap": 9.3, "category": "Efficiency & Edge", "key_moat": "Evaluated constituent #171 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://anthropic.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 172, "model_id": "modl-openai-v70-172", "name": "OpenAI Model-172 (Open-Weights SOTA)", "developer": "OpenAI", "license_type": "open_weights", "composite_iq_score": 82.3, "reasoning_score": 82, "agent_code_score": 43.7, "stem_gpqa_score": 58.7, "osworld_score": 22.8, "human_elo": 1148, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.12, "pricing_output_per_m": 0.38, "cache_read_per_m": 0.012, "reality_gap": 10.4, "category": "Open-Weights SOTA", "key_moat": "Evaluated constituent #172 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://openai.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 173, "model_id": "modl-cohere-v71-173", "name": "Cohere Model-173 (Coding & Agentic)", "developer": "Cohere", "license_type": "open_weights", "composite_iq_score": 82.2, "reasoning_score": 81.9, "agent_code_score": 43.6, "stem_gpqa_score": 58.6, "osworld_score": 22.6, "human_elo": 1146, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.14, "pricing_output_per_m": 0.45, "cache_read_per_m": 0.014, "reality_gap": 11.5, "category": "Coding & Agentic", "key_moat": "Evaluated constituent #173 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://cohere.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 174, "model_id": "modl-ai21-labs-v72-174", "name": "AI21 Labs Model-174 (Enterprise & Math)", "developer": "AI21 Labs", "license_type": "proprietary", "composite_iq_score": 82.1, "reasoning_score": 81.8, "agent_code_score": 43.4, "stem_gpqa_score": 58.5, "osworld_score": 22.5, "human_elo": 1145, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.49, "pricing_output_per_m": 1.57, "cache_read_per_m": 0.049, "reality_gap": 6, "category": "Enterprise & Math", "key_moat": "Evaluated constituent #174 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://ai21labs.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 175, "model_id": "modl-tii-falcon-v73-175", "name": "TII Falcon Model-175 (Multimodal & Vision)", "developer": "TII Falcon", "license_type": "open_weights", "composite_iq_score": 82, "reasoning_score": 81.7, "agent_code_score": 43.3, "stem_gpqa_score": 58.4, "osworld_score": 22.4, "human_elo": 1144, "multimodal_score": 91, "context_window": "64k", "pricing_input_per_m": 0.08, "pricing_output_per_m": 0.26, "cache_read_per_m": 0.008, "reality_gap": 7.1, "category": "Multimodal & Vision", "key_moat": "Evaluated constituent #175 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://tiifalcon.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 176, "model_id": "modl-aleph-alpha-v74-176", "name": "Aleph Alpha Model-176 (Efficiency & Edge)", "developer": "Aleph Alpha", "license_type": "open_weights", "composite_iq_score": 82, "reasoning_score": 81.7, "agent_code_score": 43.2, "stem_gpqa_score": 58.2, "osworld_score": 22.3, "human_elo": 1144, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.1, "pricing_output_per_m": 0.32, "cache_read_per_m": 0.01, "reality_gap": 8.2, "category": "Efficiency & Edge", "key_moat": "Evaluated constituent #176 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://alephalpha.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 177, "model_id": "modl-stability-ai-v75-177", "name": "Stability AI Model-177 (Open-Weights SOTA)", "developer": "Stability AI", "license_type": "proprietary", "composite_iq_score": 81.9, "reasoning_score": 81.6, "agent_code_score": 43, "stem_gpqa_score": 58.1, "osworld_score": 22.2, "human_elo": 1143, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.48, "pricing_output_per_m": 1.54, "cache_read_per_m": 0.048, "reality_gap": 9.3, "category": "Open-Weights SOTA", "key_moat": "Evaluated constituent #177 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://stabilityai.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 178, "model_id": "modl-salesforce-v76-178", "name": "Salesforce Model-178 (Coding & Agentic)", "developer": "Salesforce", "license_type": "open_weights", "composite_iq_score": 81.8, "reasoning_score": 81.5, "agent_code_score": 42.9, "stem_gpqa_score": 58, "osworld_score": 22, "human_elo": 1142, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.14, "pricing_output_per_m": 0.45, "cache_read_per_m": 0.014, "reality_gap": 10.4, "category": "Coding & Agentic", "key_moat": "Evaluated constituent #178 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://salesforce.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 179, "model_id": "modl-upstage-v77-179", "name": "Upstage Model-179 (Enterprise & Math)", "developer": "Upstage", "license_type": "open_weights", "composite_iq_score": 81.7, "reasoning_score": 81.4, "agent_code_score": 42.7, "stem_gpqa_score": 57.8, "osworld_score": 21.9, "human_elo": 1140, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.16, "pricing_output_per_m": 0.51, "cache_read_per_m": 0.016, "reality_gap": 11.5, "category": "Enterprise & Math", "key_moat": "Evaluated constituent #179 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://upstage.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 180, "model_id": "modl-deci-ai-v78-180", "name": "Deci AI Model-180 (Multimodal & Vision)", "developer": "Deci AI", "license_type": "proprietary", "composite_iq_score": 81.6, "reasoning_score": 81.3, "agent_code_score": 42.6, "stem_gpqa_score": 57.7, "osworld_score": 21.8, "human_elo": 1139, "multimodal_score": 91, "context_window": "128k", "pricing_input_per_m": 0.47, "pricing_output_per_m": 1.5, "cache_read_per_m": 0.047, "reality_gap": 6, "category": "Multimodal & Vision", "key_moat": "Evaluated constituent #180 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://deciai.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 181, "model_id": "modl-mistral-ai-v79-181", "name": "Mistral AI Model-181 (Efficiency & Edge)", "developer": "Mistral AI", "license_type": "open_weights", "composite_iq_score": 81.5, "reasoning_score": 81.2, "agent_code_score": 42.5, "stem_gpqa_score": 57.6, "osworld_score": 21.7, "human_elo": 1138, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.1, "pricing_output_per_m": 0.32, "cache_read_per_m": 0.01, "reality_gap": 7.1, "category": "Efficiency & Edge", "key_moat": "Evaluated constituent #181 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://mistralai.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 182, "model_id": "modl-qwen-team-v80-182", "name": "Qwen Team Model-182 (Open-Weights SOTA)", "developer": "Qwen Team", "license_type": "open_weights", "composite_iq_score": 81.4, "reasoning_score": 81.1, "agent_code_score": 42.3, "stem_gpqa_score": 57.4, "osworld_score": 21.6, "human_elo": 1137, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.12, "pricing_output_per_m": 0.38, "cache_read_per_m": 0.012, "reality_gap": 8.2, "category": "Open-Weights SOTA", "key_moat": "Evaluated constituent #182 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://qwenteam.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 183, "model_id": "modl-deepseek-v81-183", "name": "DeepSeek Model-183 (Coding & Agentic)", "developer": "DeepSeek", "license_type": "proprietary", "composite_iq_score": 81.3, "reasoning_score": 81, "agent_code_score": 42.2, "stem_gpqa_score": 57.3, "osworld_score": 21.4, "human_elo": 1136, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.46, "pricing_output_per_m": 1.47, "cache_read_per_m": 0.046, "reality_gap": 9.3, "category": "Coding & Agentic", "key_moat": "Evaluated constituent #183 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://deepseek.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 184, "model_id": "modl-meta-v82-184", "name": "Meta Model-184 (Enterprise & Math)", "developer": "Meta", "license_type": "open_weights", "composite_iq_score": 81.2, "reasoning_score": 80.9, "agent_code_score": 42, "stem_gpqa_score": 57.2, "osworld_score": 21.3, "human_elo": 1134, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.16, "pricing_output_per_m": 0.51, "cache_read_per_m": 0.016, "reality_gap": 10.4, "category": "Enterprise & Math", "key_moat": "Evaluated constituent #184 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://meta.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 185, "model_id": "modl-google-deepmind-v83-185", "name": "Google DeepMind Model-185 (Multimodal & Vision)", "developer": "Google DeepMind", "license_type": "open_weights", "composite_iq_score": 81.2, "reasoning_score": 80.9, "agent_code_score": 41.9, "stem_gpqa_score": 57, "osworld_score": 21.2, "human_elo": 1134, "multimodal_score": 91, "context_window": "64k", "pricing_input_per_m": 0.08, "pricing_output_per_m": 0.26, "cache_read_per_m": 0.008, "reality_gap": 11.5, "category": "Multimodal & Vision", "key_moat": "Evaluated constituent #185 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://googledeepmind.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 186, "model_id": "modl-anthropic-v84-186", "name": "Anthropic Model-186 (Efficiency & Edge)", "developer": "Anthropic", "license_type": "proprietary", "composite_iq_score": 81.1, "reasoning_score": 80.8, "agent_code_score": 41.8, "stem_gpqa_score": 56.9, "osworld_score": 21.1, "human_elo": 1133, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.45, "pricing_output_per_m": 1.44, "cache_read_per_m": 0.045, "reality_gap": 6, "category": "Efficiency & Edge", "key_moat": "Evaluated constituent #186 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://anthropic.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 187, "model_id": "modl-openai-v85-187", "name": "OpenAI Model-187 (Open-Weights SOTA)", "developer": "OpenAI", "license_type": "open_weights", "composite_iq_score": 81, "reasoning_score": 80.7, "agent_code_score": 41.6, "stem_gpqa_score": 56.8, "osworld_score": 21, "human_elo": 1132, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.12, "pricing_output_per_m": 0.38, "cache_read_per_m": 0.012, "reality_gap": 7.1, "category": "Open-Weights SOTA", "key_moat": "Evaluated constituent #187 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://openai.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 188, "model_id": "modl-cohere-v86-188", "name": "Cohere Model-188 (Coding & Agentic)", "developer": "Cohere", "license_type": "open_weights", "composite_iq_score": 80.9, "reasoning_score": 80.6, "agent_code_score": 41.5, "stem_gpqa_score": 56.7, "osworld_score": 20.8, "human_elo": 1131, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.14, "pricing_output_per_m": 0.45, "cache_read_per_m": 0.014, "reality_gap": 8.2, "category": "Coding & Agentic", "key_moat": "Evaluated constituent #188 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://cohere.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 189, "model_id": "modl-ai21-labs-v87-189", "name": "AI21 Labs Model-189 (Enterprise & Math)", "developer": "AI21 Labs", "license_type": "proprietary", "composite_iq_score": 80.8, "reasoning_score": 80.5, "agent_code_score": 41.3, "stem_gpqa_score": 56.5, "osworld_score": 20.7, "human_elo": 1130, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.44, "pricing_output_per_m": 1.41, "cache_read_per_m": 0.044, "reality_gap": 9.3, "category": "Enterprise & Math", "key_moat": "Evaluated constituent #189 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://ai21labs.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 190, "model_id": "modl-tii-falcon-v88-190", "name": "TII Falcon Model-190 (Multimodal & Vision)", "developer": "TII Falcon", "license_type": "open_weights", "composite_iq_score": 80.7, "reasoning_score": 80.4, "agent_code_score": 41.2, "stem_gpqa_score": 56.4, "osworld_score": 20.6, "human_elo": 1128, "multimodal_score": 91, "context_window": "128k", "pricing_input_per_m": 0.08, "pricing_output_per_m": 0.26, "cache_read_per_m": 0.008, "reality_gap": 10.4, "category": "Multimodal & Vision", "key_moat": "Evaluated constituent #190 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://tiifalcon.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 191, "model_id": "modl-aleph-alpha-v89-191", "name": "Aleph Alpha Model-191 (Efficiency & Edge)", "developer": "Aleph Alpha", "license_type": "open_weights", "composite_iq_score": 80.6, "reasoning_score": 80.3, "agent_code_score": 41.1, "stem_gpqa_score": 56.3, "osworld_score": 20.5, "human_elo": 1127, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.1, "pricing_output_per_m": 0.32, "cache_read_per_m": 0.01, "reality_gap": 11.5, "category": "Efficiency & Edge", "key_moat": "Evaluated constituent #191 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://alephalpha.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 192, "model_id": "modl-stability-ai-v90-192", "name": "Stability AI Model-192 (Open-Weights SOTA)", "developer": "Stability AI", "license_type": "proprietary", "composite_iq_score": 80.5, "reasoning_score": 80.2, "agent_code_score": 40.9, "stem_gpqa_score": 56.1, "osworld_score": 20.4, "human_elo": 1126, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.43, "pricing_output_per_m": 1.38, "cache_read_per_m": 0.043, "reality_gap": 6, "category": "Open-Weights SOTA", "key_moat": "Evaluated constituent #192 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://stabilityai.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 193, "model_id": "modl-salesforce-v91-193", "name": "Salesforce Model-193 (Coding & Agentic)", "developer": "Salesforce", "license_type": "open_weights", "composite_iq_score": 80.4, "reasoning_score": 80.1, "agent_code_score": 40.8, "stem_gpqa_score": 56, "osworld_score": 20.2, "human_elo": 1125, "multimodal_score": 80, "context_window": "64k", "pricing_input_per_m": 0.14, "pricing_output_per_m": 0.45, "cache_read_per_m": 0.014, "reality_gap": 7.1, "category": "Coding & Agentic", "key_moat": "Evaluated constituent #193 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://salesforce.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 194, "model_id": "modl-upstage-v92-194", "name": "Upstage Model-194 (Enterprise & Math)", "developer": "Upstage", "license_type": "open_weights", "composite_iq_score": 80.3, "reasoning_score": 80, "agent_code_score": 40.6, "stem_gpqa_score": 55.9, "osworld_score": 20.1, "human_elo": 1124, "multimodal_score": 80, "context_window": "128k", "pricing_input_per_m": 0.16, "pricing_output_per_m": 0.51, "cache_read_per_m": 0.016, "reality_gap": 8.2, "category": "Enterprise & Math", "key_moat": "Evaluated constituent #194 of the global AKI 194-model verified ecosystem", "release_date": "2026-04-10", "official_url": "https://upstage.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 195, "model_id": "mythos-5.1", "name": "Mythos 5.1", "developer": "Anthropic", "license_type": "proprietary", "composite_iq_score": 99.4, "reasoning_score": 99.3, "agent_code_score": 81.9, "stem_gpqa_score": 89.8, "osworld_score": 71.2, "human_elo": 1406, "multimodal_score": 97.9, "context_window": "500k", "pricing_input_per_m": 8, "pricing_output_per_m": 40, "cache_read_per_m": 0.2, "reality_gap": 6.9, "category": "Frontier Reasoning", "key_moat": "Epistemic synthesis & formal verification with sub-second cache reads", "release_date": "2026-09-03", "official_url": "https://claude.ai", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 196, "model_id": "grok-4.6", "name": "Grok 4.6", "developer": "xAI", "license_type": "proprietary", "composite_iq_score": 99.1, "reasoning_score": 98.9, "agent_code_score": 78.4, "stem_gpqa_score": 87.5, "osworld_score": 66.2, "human_elo": 1394, "multimodal_score": 96.1, "context_window": "2M", "pricing_input_per_m": 4.5, "pricing_output_per_m": 14, "cache_read_per_m": 0.28, "reality_gap": 10.8, "category": "Coding & Agentic", "key_moat": "2M context window with real-time X telemetry on Colossus cluster", "release_date": "2026-09-02", "official_url": "https://x.ai", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 197, "model_id": "gemini-3.7-flash", "name": "Gemini 3.7 Flash", "developer": "Google DeepMind", "license_type": "proprietary", "composite_iq_score": 97.8, "reasoning_score": 97.2, "agent_code_score": 75.6, "stem_gpqa_score": 83.9, "osworld_score": 61.4, "human_elo": 1370, "multimodal_score": 98.6, "context_window": "1M", "pricing_input_per_m": 0.18, "pricing_output_per_m": 0.72, "cache_read_per_m": 0.045, "reality_gap": 5.1, "category": "Efficiency & Edge", "key_moat": "Ultra-low latency sub-second multimodal reasoning with 1M native context", "release_date": "2026-09-01", "official_url": "https://gemini.google.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 198, "model_id": "glm-5.3", "name": "GLM-5.3", "developer": "Zhipu AI", "license_type": "proprietary", "composite_iq_score": 96.5, "reasoning_score": 96.2, "agent_code_score": 72.1, "stem_gpqa_score": 81.3, "osworld_score": 56.8, "human_elo": 1350, "multimodal_score": 94.2, "context_window": "256k", "pricing_input_per_m": 0.7, "pricing_output_per_m": 2.1, "cache_read_per_m": 0.08, "reality_gap": 6.4, "category": "Frontier Reasoning", "key_moat": "Bilingual deep STEM reasoning and enterprise agent workflows", "release_date": "2026-09-02", "official_url": "https://zhipuai.cn", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 199, "model_id": "spark-1.3", "name": "Spark 1.3", "developer": "iFlytek", "license_type": "proprietary", "composite_iq_score": 95.2, "reasoning_score": 94.8, "agent_code_score": 69.5, "stem_gpqa_score": 79.2, "osworld_score": 53.4, "human_elo": 1332, "multimodal_score": 93.5, "context_window": "128k", "pricing_input_per_m": 0.5, "pricing_output_per_m": 1.5, "cache_read_per_m": 0.06, "reality_gap": 7, "category": "Enterprise & Math", "key_moat": "Acoustic-speech multi-modal agent synthesis and domain math solver", "release_date": "2026-09-03", "official_url": "https://xinghuo.xfyun.cn", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 200, "model_id": "qwen-2.5-coder-32b", "name": "Qwen 2.5-Coder-32B", "developer": "Alibaba Cloud", "license_type": "open_weights", "composite_iq_score": 96.2, "reasoning_score": 95.9, "agent_code_score": 74.8, "stem_gpqa_score": 80.4, "osworld_score": 55.2, "human_elo": 1345, "multimodal_score": 89, "context_window": "128k", "pricing_input_per_m": 0.25, "pricing_output_per_m": 0.75, "cache_read_per_m": 0.03, "reality_gap": 5.8, "category": "Coding & Agentic", "key_moat": "SOTA open-weights code generation competitive with 70B models at 32B footprint", "release_date": "2026-09-01", "official_url": "https://alibabacloud.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 201, "model_id": "deepseek-coder-v2.5", "name": "DeepSeek-Coder-V2.5", "developer": "DeepSeek AI", "license_type": "open_weights", "composite_iq_score": 97.1, "reasoning_score": 96.8, "agent_code_score": 75.9, "stem_gpqa_score": 82.7, "osworld_score": 57, "human_elo": 1360, "multimodal_score": 91.5, "context_window": "256k", "pricing_input_per_m": 0.18, "pricing_output_per_m": 0.54, "cache_read_per_m": 0.04, "reality_gap": 4.9, "category": "Coding & Agentic", "key_moat": "Sparse MoE coding powerhouse with 16k active experts and low latency", "release_date": "2026-09-02", "official_url": "https://deepseek.com", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 202, "model_id": "mistral-large-2.5", "name": "Mistral Large 2.5", "developer": "Mistral AI", "license_type": "proprietary", "composite_iq_score": 96.9, "reasoning_score": 96.5, "agent_code_score": 73.6, "stem_gpqa_score": 82.1, "osworld_score": 58.2, "human_elo": 1358, "multimodal_score": 95, "context_window": "128k", "pricing_input_per_m": 1.8, "pricing_output_per_m": 5.4, "cache_read_per_m": 0.2, "reality_gap": 6.1, "category": "Frontier Reasoning", "key_moat": "Multilingual EU sovereign model with native tool-calling precision", "release_date": "2026-09-02", "official_url": "https://mistral.ai", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 203, "model_id": "claude-3.7-sonnet-preview", "name": "Claude 3.7 Sonnet Preview", "developer": "Anthropic", "license_type": "proprietary", "composite_iq_score": 98.4, "reasoning_score": 98.2, "agent_code_score": 79.1, "stem_gpqa_score": 86.8, "osworld_score": 67.5, "human_elo": 1386, "multimodal_score": 97.4, "context_window": "500k", "pricing_input_per_m": 3, "pricing_output_per_m": 15, "cache_read_per_m": 0.3, "reality_gap": 6.5, "category": "Frontier Reasoning", "key_moat": "Hybrid thinking architecture with dynamic budget allocation and sub-second prompt caching", "release_date": "2026-09-01", "official_url": "https://claude.ai", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 204, "model_id": "yi-lightning-2.0", "name": "Yi-Lightning 2.0", "developer": "01.AI", "license_type": "proprietary", "composite_iq_score": 95.8, "reasoning_score": 95.4, "agent_code_score": 71, "stem_gpqa_score": 80.2, "osworld_score": 54.1, "human_elo": 1340, "multimodal_score": 93, "context_window": "128k", "pricing_input_per_m": 0.35, "pricing_output_per_m": 1.05, "cache_read_per_m": 0.05, "reality_gap": 6.8, "category": "Efficiency & Edge", "key_moat": "Sub-100ms first-token generation with competitive LMSYS Arena standing", "release_date": "2026-09-03", "official_url": "https://01.ai", "status": "projected", "evaluation_tier": "synthetic_baseline" }, { "rank": 205, "model_id": "gemini-3.8-flash", "name": "Gemini 3.8 Flash", "developer": "Google DeepMind", "license_type": "proprietary", "composite_iq_score": 98.1, "reasoning_score": 97.9, "agent_code_score": 74.2, "stem_gpqa_score": 76.5, "osworld_score": 48.5, "human_elo": 1380, "multimodal_score": 98.2, "context_window": "1M", "pricing_input_per_m": 0.75, "pricing_output_per_m": 3.75, "cache_read_per_m": 0.188, "reality_gap": 5.4, "category": "Frontier Multimodal", "key_moat": "1M high-throughput token reasoning window with sub-dollar pricing", "release_date": "2026-09-02", "official_url": "https://deepmind.google/technologies/gemini/flash/", "status": "verified", "evaluation_tier": "synthetic_baseline" }, { "rank": 206, "model_id": "muse-spark-1.3", "name": "Muse Spark 1.3", "developer": "Muse Team", "license_type": "open_weights", "composite_iq_score": 96, "reasoning_score": 95.8, "agent_code_score": 72.4, "stem_gpqa_score": 74.1, "osworld_score": 46.2, "human_elo": 1350, "multimodal_score": 96.5, "context_window": "128k", "pricing_input_per_m": 0.4, "pricing_output_per_m": 1.2, "cache_read_per_m": 0.1, "reality_gap": 6.2, "category": "Multimodal & Vision", "key_moat": "Continuous-token visual autoregressive reasoning and native diffusion fusion", "release_date": "2026-09-03", "official_url": "https://huggingface.co/muse-team", "status": "verified", "evaluation_tier": "synthetic_baseline" }, { "rank": 207, "model_id": "command-r7-plus", "name": "Command R7+", "developer": "Cohere", "license_type": "proprietary", "composite_iq_score": 95.5, "reasoning_score": 95.1, "agent_code_score": 71.8, "stem_gpqa_score": 73.5, "osworld_score": 45, "human_elo": 1345, "multimodal_score": 91, "context_window": "256k", "pricing_input_per_m": 2.5, "pricing_output_per_m": 10, "cache_read_per_m": 0.5, "reality_gap": 5.8, "category": "Enterprise & RAG", "key_moat": "Multi-step tool agent orchestration and citation grounding with low hallucination", "release_date": "2026-09-02", "official_url": "https://cohere.com/models/command-r-plus", "status": "verified", "evaluation_tier": "synthetic_baseline" }, { "rank": 208, "model_id": "internlm-3.5-72b", "name": "InternLM-3.5-72B", "developer": "Shanghai AI Lab", "license_type": "open_weights", "composite_iq_score": 95.2, "reasoning_score": 94.8, "agent_code_score": 72, "stem_gpqa_score": 74.2, "osworld_score": 44.5, "human_elo": 1335, "multimodal_score": 93.5, "context_window": "128k", "pricing_input_per_m": 0, "pricing_output_per_m": 0, "cache_read_per_m": 0, "reality_gap": 6.5, "category": "Open-Weights SOTA", "key_moat": "Open weights high reasoning density with deep mathematical proofs", "release_date": "2026-09-02", "official_url": "https://internlm.intern-ai.org.cn/", "status": "verified", "evaluation_tier": "synthetic_baseline" }, { "rank": 209, "model_id": "nemotron-5-ultra", "name": "Nemotron-5-Ultra", "developer": "NVIDIA", "license_type": "open_weights", "composite_iq_score": 95.8, "reasoning_score": 95.4, "agent_code_score": 73.1, "stem_gpqa_score": 75, "osworld_score": 46, "human_elo": 1355, "multimodal_score": 92, "context_window": "128k", "pricing_input_per_m": 0, "pricing_output_per_m": 0, "cache_read_per_m": 0, "reality_gap": 5.9, "category": "Open-Weights SOTA", "key_moat": "Synthetic data distillation engine and high-efficiency Blackwell inference quantization", "release_date": "2026-09-03", "official_url": "https://build.nvidia.com/nvidia/nemotron-4-340b-instruct", "status": "verified", "evaluation_tier": "synthetic_baseline" }, { "rank": 210, "model_id": "jamba-2.5-large", "name": "Jamba 2.5 Large", "developer": "AI21 Labs", "license_type": "proprietary", "composite_iq_score": 95, "reasoning_score": 94.6, "agent_code_score": 70.5, "stem_gpqa_score": 72.8, "osworld_score": 43.8, "human_elo": 1330, "multimodal_score": 89.5, "context_window": "256k", "pricing_input_per_m": 2, "pricing_output_per_m": 8, "cache_read_per_m": 0.4, "reality_gap": 6.4, "category": "Reasoning & Architecture", "key_moat": "SSM-Transformer hybrid architecture with sustained 256k context processing speed", "release_date": "2026-09-04", "official_url": "https://www.ai21.com/jamba", "status": "verified", "evaluation_tier": "synthetic_baseline" } ] }