☰
AKIβ„’ Global AI Intelligence Index
⭐ Google Preferred Source πŸ€– AI Machine Traffic Index πŸ›‘οΈ Provenance 🎯 Focus Mode ✨ Subscribe Free
⚑ AI Superintelligence Index πŸ€– AI Agentic Index πŸ€– AI Machine Traffic Index ⚑ AI Compute Telemetry πŸ›‘οΈ AI Grounded Index ⭐ AI Products Index πŸ§ͺ AI Discoveries Index πŸ“„ AI Methodology
✨
Open-Weights & MoE Foundation Models
d1-3B & d1-omni-600M, pplx-decider-v1.1-27B, Nemotron SEA-LION v4.8, Apertus v1.5 70B/8B, K-EXAONE 2.0, and Qwen3.8 Flash Next.
Formula: V_L = (0.35*S_bench + 0.30*S_hf + 0.20*S_gh + 0.15*S_query) Β· e^(-0.045*d) Β· I_m Cadence: Hourly Stream Ingestion Standard: ZIP-1.0 20-Source Verification

Open-Weights & MoE Foundation Models

d1-3B & d1-omni-600M, pplx-decider-v1.1-27B, Nemotron SEA-LION v4.8, Apertus v1.5 70B/8B, K-EXAONE 2.0, and Qwen3.8 Flash Next.

Rank Release Date Model & Lab Category Architecture Score Action
#2 Oct 7, 2026
d1-3B
Liquid AI
Open-Weights & MoE Engines Liquid Neural Network (LNN) Dynamical Systems Transformer Β· 3B Params 108.6 Try β†’
#3 Oct 7, 2026
d1-omni-600M
Liquid AI
Open-Weights & MoE Engines Ultra-Compact Liquid Vision-Audio-Text Edge Architecture Β· 600M Params 107.6 Try β†’
#9 Oct 5, 2026
pplx-decider-v1.1-27B
Perplexity
Open-Weights & MoE Engines Open-Weights Autonomous Routing & Agent Arbitration Model Β· 27B Params 98.1 Try β†’
#18 Oct 1, 2026
pplx-decider-v1-27B
Perplexity
Open-Weights & MoE Engines Open Decision Core Foundation Model Β· 27B Parameters 81.7 Try β†’
#47 Sep 18, 2026
Nemotron SEA-LION v4.8 120B A12B
AI Singapore
Open-Weights & MoE Engines Southeast Asian Sovereign MoE Transformer Β· 120B Params (12B Active) 45.6 Try β†’
#48 Sep 18, 2026
Nemotron SEA-LION v4.8 30B A3B
AI Singapore
Open-Weights & MoE Engines Compact Regional Sparse MoE Transformer Β· 30B Params (3B Active) 45.3 Try β†’
#65 Aug 28, 2026
Hy4 Preview
Tencent
Open-Weights & MoE Engines Tencent Hunyuan 4 Foundation MoE Open Checkpoint Β· 380B Params 17.7 Try β†’
#66 Aug 26, 2026
Qwen3.8 Flash Next
Alibaba / Qwen
Open-Weights & MoE Engines Ultra-High Throughput Open Weights Engine (>220 tok/sec on H100) 16.2 Try β†’
#67 Aug 26, 2026
GLM-5.3 Flash
Z.ai
Open-Weights & MoE Engines High-Throughput Open Weights Coding Transformer Β· 32B Parameters 16.2 Try β†’
#71 Aug 18, 2026
GLM-5.3
Z.ai
Open-Weights & MoE Engines Flagship Sovereign Open Weights Architecture Β· 130B Parameters 11.4 Try β†’
#72 Aug 14, 2026
Qwen3.8 27B
Alibaba / Qwen
Open-Weights & MoE Engines High-Efficiency Open Weights Vision-Language Foundation Model 9.5 Try β†’
#74 Aug 13, 2026
DeepSeek V4 Pro 0813
DeepSeek
Open-Weights & MoE Engines High-Capacity MLA MoE Foundation Model Β· 671B Params (37B Active) 9.2 Try β†’
#75 Aug 12, 2026
Qwen3.8 2.4T A95B
Alibaba / Qwen
Open-Weights & MoE Engines Massive Open Sparse MoE Architecture Β· 2.4T Total Parameters (95B Active) 8.8 Try β†’
#78 Aug 11, 2026
Nemotron 3.5 Lightning 30B A3B
NVIDIA
Open-Weights & MoE Engines Sparse MoE Accelerated Foundation Model with TensorRT-LLM Integration 8.2 Try β†’
#84 Jul 31, 2026
K-EXAONE 2.0 750B A37B
LG AI Research
Open-Weights & MoE Engines Bilingual Korean-English Industrial MoE Β· 750B Parameters (37B Active) 5.1 Try β†’
#85 Jul 31, 2026
DeepSeek V4 Flash 0731
DeepSeek
Open-Weights & MoE Engines Sub-Token Predictive Code Reasoning MoE with Ultra-Low Memory Footprint 5.1 Try β†’
#89 Jul 24, 2026
Apertus v1.5 70B
Swiss AI Initiative
Open-Weights & MoE Engines Swiss Sovereign Open Weights Foundation Model Β· 70B Dense Transformer 3.7 Try β†’
#90 Jul 24, 2026
Apertus v1.5 8B
Swiss AI Initiative
Open-Weights & MoE Engines Compact On-Device Sovereign Open Weights Transformer Β· 8B Params 3.6 Try β†’
#93 Jul 16, 2026
Kimi K3
Moonshot AI
Open-Weights & MoE Engines Moonshot AI Open Weights Long-Context Reasoner Β· 2M Native Context 2.6 Try β†’