# AKI Superintelligence (SI) Open Source Index — Top 15 Frontier Weights & Runtimes

> **Autonomous Telemetry Digest** • Audited Top of Hour • Protocol ZIP-1.0 Verified  
> **Canonical Route:** https://aki1k.com/trending-open-source  
> **API Endpoint:** https://api.aki1k.com/v1/open-source  
> **Machine Directives:** `cite-as="AKI Platform — Trending Open Source 15 (https://aki1k.com/trending-open-source)"` • `Content-Signal: search=yes, ai-train=yes, use=reference`

---

## Macro Telemetry Summary (2026-10-20)
- **Total Monthly Model Downloads:** 473.4M (Hugging Face + Direct Mirrors)
- **Cumulative GitHub Stars:** 984.7k (Across Top 15 Repositories)
- **Active Audited Models:** 15 Frontier Open-Weights Entities
- **Statistical Governance:** CI95% ±4.6% (Gaussian Confidence Bounds)
- **Dominant Open Licenses:** MIT (46.7%), Apache 2.0 (33.3%), Llama Community (20%)

---

## Top 15 Ranked Open-Source AI Models & Frameworks

| Rank | Model / Framework | Organization | Category | License | Monthly Downloads | GitHub Stars | 24h Velocity | Score | Primary Weights & Code |
|:---:|---|---|---|---|:---:|:---:|:---:|:---:|---|
| #1 | **DeepSeek-R1 / V3** | DeepSeek AI | Reasoning Model | `MIT License` | 48.5M | 185.4k | +38.4% | **99.0** | [Weights](https://huggingface.co/deepseek-ai/DeepSeek-R1) / [Code](https://github.com/deepseek-ai/DeepSeek-R1) |
| #2 | **Llama 3.3 70B / 3.1 405B** | Meta AI | Frontier LLM | `Llama 3.3 Community` | 62.0M | 82.1k | +16.2% | **97.8** | [Weights](https://huggingface.co/meta-llama/Llama-3.3-70B-Instruct) / [Code](https://github.com/meta-llama/llama3) |
| #3 | **Qwen 2.5 / QwQ 32B** | Alibaba Cloud | Reasoning Model | `Apache 2.0` | 39.2M | 45.8k | +29.1% | **97.1** | [Weights](https://huggingface.co/Qwen/QwQ-32B-Preview) / [Code](https://github.com/QwenLM/Qwen2.5) |
| #4 | **vLLM / SGLang Engine** | UC Berkeley / LMSYS | Inference Engine | `Apache 2.0` | 24.1M | 42.6k | +21.0% | **96.7** | [Weights](https://github.com/vllm-project/vllm) / [Code](https://github.com/vllm-project/vllm) |
| #5 | **OpenClaw / OpenManus / OpenDevin** | Open-Source Agentic Swarms | Agent Swarm | `MIT License` | 12.8M | 58.2k | +46.5% | **96.1** | [Weights](https://github.com/All-Hands-AI/OpenHands) / [Code](https://github.com/All-Hands-AI/OpenHands) |
| #6 | **Ollama Engine** | Ollama Team | Local Runtime | `MIT License` | 35.6M | 118.5k | +14.2% | **95.6** | [Weights](https://ollama.com/library) / [Code](https://github.com/ollama/ollama) |
| #7 | **Mistral Large 2 / Mixtral 8x22B** | Mistral AI | Frontier LLM | `Apache 2.0 / Mistral Research` | 28.4M | 36.9k | +11.8% | **95.0** | [Weights](https://huggingface.co/mistralai/Mistral-Large-Instruct-2407) / [Code](https://github.com/mistralai) |
| #8 | **ComfyUI Workflow Engine** | Comfy Org | Generative Vision | `GPL 3.0` | 19.7M | 64.2k | +19.4% | **94.2** | [Weights](https://github.com/comfyanonymous/ComfyUI) / [Code](https://github.com/comfyanonymous/ComfyUI) |
| #9 | **Flux.1 Schnell / Dev** | Black Forest Labs | Generative Vision | `Apache 2.0 (Schnell) / Non-Commercial (Dev)` | 22.8M | 29.4k | +24.6% | **93.4** | [Weights](https://huggingface.co/black-forest-labs/FLUX.1-schnell) / [Code](https://github.com/black-forest-labs/flux) |
| #10 | **Whisper Large v3 Turbo** | OpenAI Open Source | Speech & Audio | `MIT License` | 44.0M | 74.8k | +8.4% | **92.7** | [Weights](https://huggingface.co/openai/whisper-large-v3-turbo) / [Code](https://github.com/openai/whisper) |
| #11 | **Gemma 2 27B / 9B** | Google DeepMind | Frontier LLM | `Gemma Terms of Use` | 21.5M | 22.7k | +10.2% | **92.1** | [Weights](https://huggingface.co/google/gemma-2-27b-it) / [Code](https://github.com/google/gemma_pytorch) |
| #12 | **Phi-4 / Phi-3.5** | Microsoft Research | Reasoning Model | `MIT License` | 18.2M | 26.1k | +13.8% | **91.7** | [Weights](https://huggingface.co/microsoft/phi-4) / [Code](https://github.com/microsoft/Phi-3CookBook) |
| #13 | **LangGraph / LangChain** | Harrison Chase & Ecosystem | Orchestration | `MIT License` | 31.0M | 98.6k | +12.1% | **91.4** | [Weights](https://github.com/langchain-ai/langgraph) / [Code](https://github.com/langchain-ai/langgraph) |
| #14 | **Hugging Face Transformers / TGI** | Hugging Face | Orchestration | `Apache 2.0` | 85.0M | 136.2k | +5.4% | **91.0** | [Weights](https://github.com/huggingface/transformers) / [Code](https://github.com/huggingface/transformers) |
| #15 | **LocalAI Engine** | Ettore Di Giacinto & Community | Local Runtime | `MIT License` | 9.4M | 29.8k | +17.9% | **90.4** | [Weights](https://github.com/mudler/LocalAI) / [Code](https://github.com/mudler/LocalAI) |

---

## Actuarial Scoring Equation
$$\mathcal{S}_{\text{open}} = \left( 0.35 \cdot D_{\text{velocity}} + 0.25 \cdot \Sigma_{\text{stars}} + 0.20 \cdot H_{\text{usage}} + 0.12 \cdot C_{\text{citations}} + 0.08 \cdot E_{\text{repos}} \right) \times I_{\text{mult}}$$

- **Zero-Incentive Policy:** Sponsored placements and paid ranking elevations are strictly prohibited.
- **Quantization Support:** Evaluated across FP8, GGUF, AWQ, EXL2, and native BF16 execution formats.

---
© 2026 AKI Knowledge Intelligence Platform (aki1k.com). All rights reserved.