# Open-Source LLM Leaderboard 2026

> Compare open-weight LLMs by benchmark score, license, size, context, quantization, and deployment needs.

This page is the canonical open-weight ranking. It uses the same public BenchAlign v5 overall lane as the main leaderboard, then filters to downloadable model weights. The score compares measured capability; it does not decide whether a license is permissive, a model fits your hardware, or self-hosting beats an API on cost. Use the linked decision guide for those deployment questions.

Canonical page: https://benchlm.ai/best/open-source

Last updated: September 10, 2026

## Rankings

| Rank | Model | Creator | Type | Context | Score |
|------|-------|---------|------|---------|-------|
| 1 | [Qwen3.8 Max](/models/qwen3-8-max) | Alibaba | Open Weight | 1M | 71.61 |
| 2 | [GLM-5.3](/models/glm-5-3) | Z.AI | Open Weight | 1M | 68.35 |
| 3 | [GLM-5.2](/models/glm-5-2) | Z.AI | Open Weight | 1M | 68.1 |
| 4 | [GLM-5.3-Flash](/models/glm-5-3-flash) | Z.AI | Open Weight | 1M | 66.04 |
| 5 | [Kimi K2.7 Code](/models/kimi-k2-7-code) | Moonshot AI | Open Weight | 256K | 65.52 |
| 6 | [Kimi K2.6](/models/kimi-2-6) | Moonshot AI | Open Weight | 256K | 65.42 |
| 7 | [Ornith-1.5-397B](/models/ornith-1-5-397b) | Ornith AI | Open Weight | 262K | 65.33 |
| 8 | [Qwen3.8-27B](/models/qwen3-8-27b) | Alibaba | Open Weight | 262K | 64.43 |
| 9 | [dots3-note Preview](/models/dots3-note-preview) | Dots Studio | Open Weight | 512K | 64.42 |
| 10 | [GLM-5.1](/models/glm-5-1) | Z.AI | Open Weight | 203K | 63.49 |
| 11 | [MiniMax M3](/models/minimax-m3) | MiniMax | Open Weight | 1M | 61.51 |
| 12 | [GLM-5](/models/glm-5) | Z.AI | Open Weight | 200K | 61.45 |
| 13 | [Hy4 preview](/models/hy4-preview) | Tencent | Open Weight | 1M | 61.03 |
| 14 | [Hy3](/models/hy3) | Tencent | Open Weight | 256K | 60.93 |
| 15 | [Inkling](/models/inkling) | Thinking Machines Lab | Open Weight | 1M | 60.29 |
| 16 | [Inkling-Small](/models/inkling-small) | Thinking Machines Lab | Open Weight | 1M | 59.26 |
| 17 | [Qwen3.8-Flash-Next](/models/qwen3-8-flash-next) | Alibaba | Open Weight | 262K | 58.55 |
| 18 | [GLM-5 (Reasoning)](/models/glm-5-reasoning) | Z.AI | Open Weight | 200K | 57.96 |
| 19 | [GLM-4.7](/models/glm-4-7) | Z.AI | Open Weight | 200K | 57.72 |
| 20 | [Qwen3.5 397B (Reasoning)](/models/qwen3-5-397b-reasoning) | Alibaba | Open Weight | 128K | 57.69 |
| 21 | [DeepSeek V3.2](/models/deepseek-v3-2) | DeepSeek | Open Weight | 128K | 56.91 |
| 22 | [Qwen3.5-122B-A10B](/models/qwen3-5-122b-a10b) | Alibaba | Open Weight | 262K | 56.4 |
| 23 | [DeepSeek V3.2 (Thinking)](/models/deepseek-v3-2-thinking) | DeepSeek | Open Weight | 128K | 56.34 |
| 24 | [Qwen3 235B 2507 (Reasoning)](/models/qwen3-235b-2507-reasoning) | Alibaba | Open Weight | 128K | 56.23 |
| 25 | [Hy3 Preview](/models/hy3-preview) | Tencent | Open Weight | 256K | 55.88 |
| 26 | [Qwen3.5-27B](/models/qwen3-5-27b) | Alibaba | Open Weight | 262K | 55.31 |
| 27 | [Qwen3.5 397B](/models/qwen3-5-397b) | Alibaba | Open Weight | 128K | 55.24 |
| 28 | [MiniMax M2.7](/models/minimax-m2-7) | MiniMax | Open Weight | 200K | 55.18 |
| 29 | [Gemma 4 26B A4B](/models/gemma-4-26b-a4b) | Google | Open Weight | 256K | 55.11 |
| 30 | [Qwen3 235B 2507](/models/qwen3-235b-2507) | Alibaba | Open Weight | 128K | 54.27 |
| 31 | [Trinity-Large-Preview](/models/trinity-large-preview) | Arcee AI | Open Weight | 512K | 54.2 |
| 32 | [Kimi K2.5](/models/kimi-k2-5) | Moonshot AI | Open Weight | 256K | 54.02 |
| 33 | [Qwen3.5-35B-A3B](/models/qwen3-5-35b-a3b) | Alibaba | Open Weight | 262K | 53.41 |
| 34 | [DeepSeek LLM 2.0](/models/deepseek-llm-2-0) | DeepSeek | Open Weight | 128K | 52.81 |
| 35 | [Gemma 4 31B](/models/gemma-4-31b) | Google | Open Weight | 256K | 52.57 |
| 36 | [GLM-4.6](/models/glm-4-6) | Z.AI | Open Weight | 200K | 51.86 |
| 37 | [Nemotron 3 Nano 30B](/models/nemotron-3-nano-30b) | NVIDIA | Open Weight | 32K | 51.25 |
| 38 | [DeepSeek V3.1](/models/deepseek-v3-1) | DeepSeek | Open Weight | 128K | 50.64 |
| 39 | [DeepSeek-R1](/models/deepseek-r1) | DeepSeek | Open Weight | 128K | 50.24 |
| 40 | [Step 3.7 Flash](/models/step-3-7-flash) | StepFun | Open Weight | 256K | 49.96 |
| 41 | [Step 3.5 Flash](/models/step-3-5-flash) | StepFun | Open Weight | 256K | 49.86 |
| 42 | [Nemotron 3 Super 120B A12B](/models/nemotron-3-super-120b-a12b) | NVIDIA | Open Weight | 256K | 49.32 |
| 43 | [MiMo-V2-Flash](/models/mimo-v2-flash) | Xiaomi | Open Weight | 256K | 48.99 |
| 44 | [DeepSeek V3.1 (Reasoning)](/models/deepseek-v3-1-reasoning) | DeepSeek | Open Weight | 128K | 48.82 |
| 45 | [DeepSeek Coder 2.0](/models/deepseek-coder-2-0) | DeepSeek | Open Weight | 128K | 48.64 |
| 46 | [Nemotron 3 Super 100B](/models/nemotron-3-super-100b) | NVIDIA | Open Weight | 1M | 48.45 |
| 47 | [Qwen2.5-1M](/models/qwen2-5-1m) | Alibaba | Open Weight | 1M | 48.25 |
| 48 | [DeepSeekMath V2](/models/deepseekmath-v2) | DeepSeek | Open Weight | 128K | 48.25 |
| 49 | [Qwen3.6-27B](/models/qwen3-6-27b) | Alibaba | Open Weight | 262K | 47.8 |
| 50 | [Ministral 3 14B (Reasoning)](/models/ministral-3-14b-reasoning) | Mistral | Open Weight | 128K | 47.73 |
| 51 | [Ling 3.0 Flash](/models/ling-3-0-flash) | InclusionAI | Open Weight | 262K | 47.42 |
| 52 | [GLM-4.7-Flash](/models/glm-4-7-flash) | Z.AI | Open Weight | 200K | 46.3 |
| 53 | [Qwen3.6-35B-A3B](/models/qwen3-6-35b-a3b) | Alibaba | Open Weight | 262K | 46.21 |
| 54 | [GPT-OSS 120B](/models/gpt-oss-120b) | OpenAI | Open Weight | 128K | 45.73 |
| 55 | [Muse Glimmer 30B](/models/muse-glimmer-30b) | Meta | Open Weight | 131K | 45.34 |
| 56 | [Mistral 8x7B](/models/mistral-8x7b) | Mistral | Open Weight | 32K | 43.46 |
| 57 | [Trinity-Large-Thinking](/models/trinity-large-thinking) | Arcee AI | Open Weight | 512K | 43.44 |
| 58 | [Gemma 4 12B](/models/gemma-4-12b) | Google | Open Weight | 256K | 43.25 |
| 59 | [Command A+](/models/command-a-plus) | Cohere | Open Weight | 128K | 43.11 |
| 60 | [Mistral Small 4](/models/mistral-small-4) | Mistral | Open Weight | 256K | 43.01 |
| 61 | [Nemotron Ultra 253B](/models/nemotron-ultra-253b) | NVIDIA | Open Weight | 32K | 42.89 |
| 62 | [DeepSeek V3](/models/deepseek-v3) | DeepSeek | Open Weight | 128K | 41.54 |
| 63 | [Nemotron 3 Nano Omni 30B A3B](/models/nemotron-3-nano-omni-30b-a3b) | NVIDIA | Open Weight | 256K | 41.15 |
| 64 | [Ling 2.6 Flash](/models/ling-2-6-flash) | InclusionAI | Open Weight | 262K | 40.84 |
| 65 | [GPT-OSS 20B](/models/gpt-oss-20b) | OpenAI | Open Weight | 128K | 40.78 |
| 66 | [Nemotron 3 Ultra](/models/nemotron-3-ultra) | NVIDIA | Open Weight | 1M | 40.57 |
| 67 | [Gemma 4 E4B](/models/gemma-4-e4b) | Google | Open Weight | 128K | 40.44 |
| 68 | [Sarvam 105B](/models/sarvam-105b) | Sarvam | Open Weight | 128K | 40.37 |
| 69 | [LFM2.5-2.6B](/models/lfm2-5-2-6b) | LiquidAI | Open Weight | 128K | 40.16 |
| 70 | [Gemma 4 E2B](/models/gemma-4-e2b) | Google | Open Weight | 128K | 39.83 |
| 71 | [LFM2.5-8B-A1B](/models/lfm2-5-8b-a1b) | LiquidAI | Open Weight | 128K | 39.53 |
| 72 | [Sarvam 30B](/models/sarvam-30b) | Sarvam | Open Weight | 64K | 39.18 |
| 73 | [Exaone 4.0 32B](/models/exaone-4-0-32b) | LG AI Research | Open Weight | 128K | 39.06 |
| 74 | [Granite 4.2 8B](/models/granite-4-2-8b) | IBM | Open Weight | 128K | 39 |
| 75 | [Ministral 3 8B (Reasoning)](/models/ministral-3-8b-reasoning) | Mistral | Open Weight | 128K | 38.96 |
| 76 | [Exaone 4.0 1.2B](/models/exaone-4-0-1-2b) | LG AI Research | Open Weight | 128K | 38.48 |
| 77 | [Granite-4.0-H-1B](/models/granite-4-0-h-1b) | IBM | Open Weight | 128K | 38.46 |
| 78 | [Qwen2.5-VL-32B](/models/qwen2-5-vl-32b) | Alibaba | Open Weight | 32K | 38.44 |
| 79 | [Llama 4 Behemoth](/models/llama-4-behemoth) | Meta | Open Weight | 32K | 38.4 |
| 80 | [Granite-4.0-1B](/models/granite-4-0-1b) | IBM | Open Weight | 128K | 38.37 |
| 81 | [Granite-4.0-350M](/models/granite-4-0-350m) | IBM | Open Weight | 32K | 38.27 |
| 82 | [Granite-4.0-H-350M](/models/granite-4-0-h-350m) | IBM | Open Weight | 32K | 38.27 |
| 83 | [Ministral 3 3B (Reasoning)](/models/ministral-3-3b-reasoning) | Mistral | Open Weight | 128K | 38.12 |
| 84 | [Mistral 8x7B v0.2](/models/mistral-8x7b-v0-2) | Mistral | Open Weight | 32K | 37.74 |
| 85 | [Qwen2.5-72B](/models/qwen2-5-72b) | Alibaba | Open Weight | 128K | 37.42 |
| 86 | [Ornith-1.5-35B-A3B](/models/ornith-1-5-35b-a3b) | Ornith AI | Open Weight | 262K | 37.02 |
| 87 | [Llama 3.1 405B](/models/llama-3-1-405b) | Meta | Open Weight | 128K | 36.59 |
| 88 | [Gemma 3 27B](/models/gemma-3-27b) | Google | Open Weight | 32K | 36.33 |
| 89 | [Llama 4 Scout](/models/llama-4-scout) | Meta | Open Weight | 10M | 33.68 |
| 90 | [Qwen2.5 Coder 32B Instruct](/models/qwen2-5-coder-32b-instruct) | Alibaba | Open Weight | 128K | 33.44 |
| 91 | [Phi-4](/models/phi-4) | Microsoft | Open Weight | 16K | 32.95 |
| 92 | [DeepSeek R1 Distill Qwen 32B](/models/deepseek-r1-distill-qwen-32b) | DeepSeek | Open Weight | 128K | 32.62 |
| 93 | [Llama 3 70B](/models/llama-3-70b) | Meta | Open Weight | 128K | 30.61 |
| 94 | [Ministral 3 14B](/models/ministral-3-14b) | Mistral | Open Weight | 128K | 30.35 |
| 95 | [Mistral Medium 3.5 128B](/models/mistral-medium-3-5-128b) | Mistral | Open Weight | 256K | 30.11 |
| 96 | [Mixtral 8x22B Instruct v0.1](/models/mixtral-8x22b-instruct-v0-1) | Mistral | Open Weight | 64K | 26.7 |
| 97 | [Llama 4 Maverick](/models/llama-4-maverick) | Meta | Open Weight | 1M | 22.17 |
| 98 | [Nemotron 3.5 Lightning 30B A3B NVFP4](/models/nemotron-3-5-lightning-30b-a3b-nvfp4) | NVIDIA | Open Weight | 1M | 21.11 |
| 99 | [Ministral 3 8B](/models/ministral-3-8b) | Mistral | Open Weight | 128K | 17.96 |
| 100 | [Ministral 3 3B](/models/ministral-3-3b) | Mistral | Open Weight | 128K | 16.13 |
| 101 | [Nemotron-4 15B](/models/nemotron-4-15b) | NVIDIA | Open Weight | 32K | 15.93 |
| 102 | [Mistral 7B v0.3](/models/mistral-7b-v0-3) | Mistral | Open Weight | 32K | 8.69 |
| 103 | [Laguna XS.2](/models/laguna-xs-2) | Poolside | Open Weight | 256K | 1.52 |

## Key Takeaways

- Top model: [Qwen3.8 Max](/models/qwen3-8-max) with a score of 71.61
- Best open-weight option: [Qwen3.8 Max](/models/qwen3-8-max) at #1
- Models included: 103

## Compare the Leaders

- [Qwen3.8 Max vs GLM-5.3](/compare/glm-5-3-vs-qwen3-8-max)