Score74.6
License
Kimi K3.1 Pro License
Community/custom
Parameters
2800B MoE (104B active)
INT4, Q2
Context1.05M
Reference hardware
8× NVIDIA B300 (288GB) · 2304GB total
Recommendation
Ranking data as of
Open-Source LLM Leaderboard 2026
The best open-source LLM by current open-weight benchmark score is Kimi K3.1 Pro. It leads the November 2026 open-weight ranking at 74.6, ahead of Kimi K3 (69.2) and Ling 3.1 Flash (65.3). The license directory below separates OSI-approved licenses from community terms.
Bottom line: use the live table above for capability order, then check evidence status, license terms, memory requirements, and serving cost before choosing a deployment.
Citable stat94 open-weight models are ranked as of November 11, 2026; Kimi K3.1 Pro leads at 74.6/100.
Full Rankings (94 ranked)
Current position
Kimi K3.1 Pro leads the live open-weight ranking at 74.6 with Estimated evidence.
Kimi K3 ranks #2 at 69.2 with Supported evidence.
Ling 3.1 Flash ranks #3 at 65.3 with Estimated evidence.
How to choose
Best open-weight model overall?The #1 row of the ranking above, then its evidence label and license termsCoding-focused open model?Use the open rows on the calibrated coding leaderboardSelf-hosting economics?Estimate hardware and API break-even before choosing a model familyChinese-lab open models?See the Chinese model rankings — most frontier open weights come from these labs
Key Takeaways
The top model is Kimi K3.1 Pro by Moonshot AI with a BenchAlign v5.8 score of 74.6 and Estimated evidence.
The best open-weight model is Kimi K3.1 Pro at position #1.
94 models are included in this ranking.
Score in Context
What these scores mean
Open-weight models are ranked by the same public BenchAlign v5.8 overall score as proprietary models. Supported and Estimated describe the evidence behind each position; they are not separate leaderboards.
Known limitations
Open weight is not the same as OSI-approved open source. The ranking does not score license restrictions, memory requirements, serving throughput, fine-tuning support, or the engineering cost of operating the model.
About this ranking
Ranking data as of November 11, 2026
This page is the canonical open-weight ranking. It uses the same public BenchAlign v5.8 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.
This is the public BenchAlign v5.8 overall lane filtered to open-weight models. The table orders point estimates. Evidence labels describe source support; conditional ranges do not establish rank confidence.
The open-weight slice starts with Kimi K3.1 Pro, followed by Kimi K3 and Ling 3.1 Flash. All rows use the public BenchAlign v5.8 projection contract. Evidence badges and score intervals matter as much as small point gaps because public source coverage is uneven.
Every ranked model publishes downloadable weights, but that does not make every license OSI-approved or every deployment practical. Use the directory below to check license terms, quantization, and reference hardware before shortlisting a model.
This ranking uses the public BenchAlign v5.8 overall contract filtered to open-weight models. For detailed model profiles, click any model name above. To compare two specific models head-to-head, use the "vs #" links.
Deployment evidence
Compare license, size, and deployment
The performance table ranks every eligible open-weight row. This smaller directory includes only models with a deployment record in the self-host catalog, so license and hardware claims remain auditable instead of being filled from family names.
Open weight is an access category, not a license verdict. “OSI-approved” appears only when the deployment catalog records MIT or Apache 2.0. Community and custom licenses stay separate, even when their weights are downloadable.
Showing 67 of 67 deployment-documented rows
Deployment catalog checked 2026-11-04.
- Score69.2License
Kimi K3 License
Community/custom
Parameters2800B MoE (104B active)
INT4, Q2
Context1.05MReference hardware8× NVIDIA B300 (288GB) · 2304GB total
- Score65.3License
MIT
OSI-approved
Parameters560B MoE (25B active)
FP8, Q4
Context262KReference hardware8× NVIDIA H100 (80GB) · 640GB total
- Score65.1License
MIT
OSI-approved
Parameters552B MoE (16B active)
FP8, Q4
Context1MReference hardware8× NVIDIA H100 (80GB) · 640GB total
- Score64.7License
Qwen3.8-Max License
Community/custom
Parameters2446B MoE (95B active)
Q4, Q2
Context1MReference hardware8× NVIDIA B300 (288GB) · 2304GB total
- Score63.8License
Apache-2.0
OSI-approved
Parameters280B MoE (16B active)
FP8, Q4
Context512KReference hardware4× NVIDIA H100 (80GB) · 320GB total
- Score62.6License
Apache-2.0
OSI-approved
Parameters35.95B MoE (3B active)
BF16, Q4
Context262KReference hardware1× NVIDIA RTX 4090 (24GB) · 24GB total
- Score62.4License
MIT
OSI-approved
Parameters1600B MoE (49B active)
Q4, Q2
Context1MReference hardware8× NVIDIA H200 (141GB) · 1128GB total
- Score62.2License
MIT
OSI-approved
Parameters397B MoE (17B active)
FP8, Q4
Context262KReference hardware4× NVIDIA H100 (80GB) · 320GB total
- Score61.2License
GLM-5.3 License
Community/custom
Parameters753B MoE (40B active)
FP8, Q4
Context1MReference hardware8× NVIDIA H100 (80GB) · 640GB total
- Score61License
Apache-2.0
OSI-approved
Parameters770B MoE (49B active)
FP8, Q4
Context1MReference hardware8× NVIDIA H100 (80GB) · 640GB total
- Score60.1License
MIT
OSI-approved
Parameters744B MoE (40B active)
FP8, Q4
Context1MReference hardware8× NVIDIA H100 (80GB) · 640GB total
- Score57.9License
MIT
OSI-approved
Parameters320B MoE (18B active)
FP8, Q4
Context1MReference hardware4× NVIDIA H100 (80GB) · 320GB total
- Score57.4License
Modified MIT
Community/custom
Parameters1000B MoE (32B active)
INT4, FP8, Q4
Context256KReference hardware8× NVIDIA H100 (80GB) · 640GB total
- Score57.2License
MIT
OSI-approved
Parameters1020B MoE (42B active)
FP8, Q4
Context1MReference hardware8× NVIDIA H200 (141GB) · 1128GB total
- Score56.7License
StepFun Model License 1.0
Community/custom
Parameters600B MoE (27B active)
FP8, Q4
Context1MReference hardware8× NVIDIA H100 (80GB) · 640GB total
- Score56License
MIT
OSI-approved
Parameters744B MoE (40B active)
FP8, Q4, Q2
Context203KReference hardware8× NVIDIA H100 (80GB) · 640GB total
- Score55.7License
Qwen Community 1.0
Community/custom
Parameters180B MoE (6B active)
FP8, Q4
Context262KReference hardware2× NVIDIA H100 (80GB) · 160GB total
- Score55.1License
MIT
OSI-approved
Parameters309B MoE (15B active)
FP8, Q4
Context1MReference hardware4× NVIDIA H100 (80GB) · 320GB total
- Score55.1License
Apache-2.0
OSI-approved
Parameters397B MoE (17B active)
FP8, Q4
Context128KReference hardware4× NVIDIA H100 (80GB) · 320GB total
- Score54.4License
MIT
OSI-approved
Parameters744B MoE (40B active)
FP8, Q4
Context200KReference hardware8× NVIDIA H100 (80GB) · 640GB total
- Score54.4License
Apache-2.0
OSI-approved
Parameters27B dense
BF16, Q4
Context262KReference hardware1× NVIDIA RTX 4090 (24GB) · 24GB total
- Score53.6License
Modified MIT
Community/custom
Parameters1000B MoE (32B active)
INT4, FP8, Q4
Context256KReference hardware8× NVIDIA H100 (80GB) · 640GB total
- Score52.7License
Modified MIT
Community/custom
Parameters1000B MoE (32B active)
INT4, FP8
Context256KReference hardware8× NVIDIA H100 (80GB) · 640GB total
- Score52.6License
Apache-2.0
OSI-approved
Parameters122B MoE (10B active)
FP8, Q4
Context262KReference hardware1× NVIDIA H100 (80GB) · 80GB total
- Score52.1License
Modified MIT
Community/custom
Parameters1050B MoE (52B active)
FP8, Q4
Context1MReference hardware8× NVIDIA H200 (141GB) · 1128GB total
- Score51.2License
Muse Model License
Community/custom
Parameters816B MoE (42B active)
FP8, Q4
Context1MReference hardware8× NVIDIA H100 (80GB) · 640GB total
- Score49.2License
OpenMDW-1.1
Community/custom
Parameters120B MoE (12B active)
FP8, Q4, NVFP4
Context1MReference hardware1× NVIDIA RTX PRO 6000 Blackwell (96GB) · 96GB total
- Score49License
MiniMax Community License
Community/custom
Parameters428B MoE (23B active)
FP8, Q4
Context1MReference hardware4× NVIDIA H100 (80GB) · 320GB total
- Score48.3License
Apache-2.0
OSI-approved
Parameters27B dense
FP8, Q4
Context262KReference hardware1× NVIDIA RTX 4090 (24GB) · 24GB total
- Score48.1License
Apache-2.0
OSI-approved
Parameters975B MoE (41B active)
FP8, Q4
Context1MReference hardware8× NVIDIA H100 (80GB) · 640GB total
- Score47.7License
MiniMax License
Community/custom
Parameters230B MoE (10B active)
FP8, Q4
Context200KReference hardware2× NVIDIA H100 (80GB) · 160GB total
- Score46.5License
OpenMDW-1.1
Community/custom
Parameters550B MoE (55B active)
FP8, Q4, NVFP4
Context1MReference hardware4× NVIDIA RTX PRO 6000 Blackwell (96GB) · 384GB total
- Score45.8License
MIT
OSI-approved
Parameters124B MoE (5.1B active)
FP8, Q4
Context262KReference hardware2× NVIDIA H100 (80GB) · 160GB total
SourceNot linked - Score45.5License
Apache-2.0
OSI-approved
Parameters35B MoE (3B active)
BF16, Q4
Context262KReference hardware1× NVIDIA RTX 4090 (24GB) · 24GB total
- Score45.4License
Apache-2.0
OSI-approved
Parameters201B MoE (11B active)
FP8, Q4, NVFP4
Context256KReference hardware2× NVIDIA RTX PRO 6000 Blackwell (96GB) · 192GB total
- Score43.8License
Apache-2.0
OSI-approved
Parameters501B MoE (23B active)
FP8, Q4, NVFP4
Context1MReference hardware4× NVIDIA RTX PRO 6000 Blackwell (96GB) · 384GB total
- Score43.8License
OpenMDW-1.1
Community/custom
Parameters398B MoE (13B active)
FP8, Q4
Context512KReference hardware4× NVIDIA H100 (80GB) · 320GB total
SourceNot linked - Score42.9License
Apache-2.0
OSI-approved
Parameters35B MoE (3B active)
BF16, Q4
Context262KReference hardware1× NVIDIA RTX 4090 (24GB) · 24GB total
- Score42.6License
OpenMDW-1.1
Community/custom
Parameters74B MoE (9B active)
FP8, Q4, INT4
Context256KReference hardware1× NVIDIA RTX 4090 (24GB) · 24GB total
- Score42.5License
Apache-2.0
OSI-approved
Parameters196B MoE (11B active)
FP8, Q4
Context256KReference hardware2× NVIDIA H100 (80GB) · 160GB total
- Score41.4License
Apache-2.0
OSI-approved
Parameters25.2B MoE (3.8B active)
BF16, Q4
Context256KReference hardware1× NVIDIA RTX 4090 (24GB) · 24GB total
- Score40.8License
MIT
OSI-approved
Parameters671B MoE (37B active)
FP8, Q4
Context128KReference hardware8× NVIDIA H100 (80GB) · 640GB total
- Score40.3License
Apache-2.0
OSI-approved
Parameters4.11B dense
BF16
Context1MReference hardware1× NVIDIA RTX 4090 (24GB) · 24GB total
- Score40.2License
MIT
OSI-approved
Parameters309B MoE (15B active)
FP8, Q4
Context256KReference hardware4× NVIDIA H100 (80GB) · 320GB total
SourceNot linked - Score38.9License
MIT
OSI-approved
Parameters31B MoE (3B active)
BF16, Q4
Context200KReference hardware1× NVIDIA RTX 4090 (24GB) · 24GB total
- Score36.8License
LFM Open License v1.0
Community/custom
Parameters24B MoE (2.45B active)
BF16, Q4
Context131KReference hardware1× NVIDIA RTX 4090 (24GB) · 24GB total
- Score34.2License
Apache-2.0
OSI-approved
Parameters675B MoE (41B active)
FP8, Q4
Context128KReference hardware8× NVIDIA H100 (80GB) · 640GB total
- Score34License
Apache-2.0
OSI-approved
Parameters8.4B MoE (0.76B active)
BF16, Q4
Context131KReference hardware1× NVIDIA RTX 4090 (24GB) · 24GB total
- Score32.6License
Apache-2.0
OSI-approved
Parameters29.6B dense
BF16, Q4
Context131KReference hardware1× NVIDIA RTX 4090 (24GB) · 24GB total
- Score32License
Apache-2.0
OSI-approved
Parameters117B MoE (5.1B active)
INT4
Context128KReference hardware1× NVIDIA H100 (80GB) · 80GB total
- Score31.7License
MIT
OSI-approved
Parameters671B MoE (37B active)
FP8, Q4
Context128KReference hardware8× NVIDIA H100 (80GB) · 640GB total
- Score30.5License
Apache-2.0
OSI-approved
Parameters106B MoE (10.3B active)
FP8, Q4
Context128KReference hardware1× NVIDIA H100 (80GB) · 80GB total
- Score29.9License
Apache-2.0
OSI-approved
Parameters218B MoE (25B active)
FP8, Q4
Context128KReference hardware2× NVIDIA H100 (80GB) · 160GB total
- Score29.7License
Modified MIT
Community/custom
Parameters128B dense
FP8, Q4
Context256KReference hardware2× NVIDIA H100 (80GB) · 160GB total
- Score29.1License
Apache-2.0
OSI-approved
Parameters31B dense
BF16, Q4
Context256KReference hardware1× NVIDIA RTX 4090 (24GB) · 24GB total
- Score28.6License
Qwen Research
Community/custom
Parameters72B dense
FP8, Q4
Context128KReference hardware2× NVIDIA RTX 4090 (24GB) · 48GB total
- Score27.9License
Llama Community
Community/custom
Parameters400B MoE (17B active)
FP8, Q4
Context1MReference hardware4× NVIDIA H100 (80GB) · 320GB total
- Score27License
Llama Community
Community/custom
Parameters109B MoE (17B active)
FP8, Q4
Context10MReference hardware1× NVIDIA H100 (80GB) · 80GB total
- Score27License
Apache-2.0
OSI-approved
Parameters119B MoE (6.5B active)
FP8, Q4
Context256KReference hardware1× NVIDIA H100 (80GB) · 80GB total
- Score26.1License
Apache-2.0
OSI-approved
Parameters8B dense
BF16
Context128KReference hardware1× NVIDIA RTX 4090 (24GB) · 24GB total
- Score25.5License
OpenMDW-1.1
Community/custom
Parameters30B MoE (3B active)
BF16, Q4, NVFP4
Context1MReference hardware1× NVIDIA RTX 4090 (24GB) · 24GB total
- Score25.3License
LFM Open License v1.0
Community/custom
Parameters8.3B MoE (1.5B active)
BF16, Q4
Context128KReference hardware1× NVIDIA RTX 4090 (24GB) · 24GB total
- Score25.2License
LFM Open License v1.0
Community/custom
Parameters2.69B dense
BF16, Q4
Context128KReference hardware1× NVIDIA RTX 4090 (24GB) · 24GB total
- Score23.6License
Apache-2.0
OSI-approved
Parameters32B MoE (2.4B active)
BF16, Q4
Context64KReference hardware1× NVIDIA RTX 4090 (24GB) · 24GB total
- Score21.4License
LFM Open License v1.0
Community/custom
Parameters1.73B dense
BF16, Q4
Context256KReference hardware1× NVIDIA RTX 4090 (24GB) · 24GB total
- Score5.3License
Apache-2.0
OSI-approved
Parameters1.08B dense
BF16, Q4
Context131KReference hardware1× NVIDIA RTX 4090 (24GB) · 24GB total
This table is the sourced deployment subset, not the complete performance ranking. A missing row means the deployment catalog is incomplete, not that the model cannot be self-hosted.
Questions
What is the best open source LLM right now?
Kimi K3.1 Pro is first by point estimate at 74.6, ahead of Kimi K3 at 69.2 and Ling 3.1 Flash at 65.3, as of November 11, 2026. The score ranges on these rows are conditional and don’t establish rank confidence, so check the evidence label beside each row before treating the order as settled.
Which LLMs are open source?
As of November 11, 2026, 94 open-weight models are ranked here. The five highest are Kimi K3.1 Pro (Moonshot AI), Kimi K3 (Moonshot AI), Ling 3.1 Flash (InclusionAI), DeepSeek V4.1 Flash 1017 (DeepSeek) and Qwen3.8 Max (Alibaba). Each publishes weights you can download. That isn’t the same as an OSI-approved open-source license, so check the license directory on this page before building on one.
Are open source LLMs as good as GPT or Claude?
Not at the top of the overall ranking, as of November 11, 2026. The highest-ranked open-weight model, Kimi K3.1 Pro, scores 74.6 and sits at #13 overall, where the first row scores 91.4. On single tasks an open model can match a closed one, and control over deployment, privacy and serving cost can still make it the better choice.
What is the best open source LLM for coding?
Use the open-weight rows on the live coding leaderboard. Coding order differs from overall order, and a single HumanEval or LiveCodeBench result should not be treated as a complete coding verdict.
Can I run these models locally?
The smaller open-weight rows run on a single consumer GPU with 4-bit quantization; frontier-size models need multi-GPU rigs or high-memory Apple Silicon. The local LLM guide breaks the rankings down by VRAM tier, and the Ollama guide includes pull commands and size estimates per model.
Which labs make the best open-weight models?
The ten highest-ranked open-weight models as of November 11, 2026 come from eight labs: Moonshot AI (2), DeepSeek (2), InclusionAI (1), Alibaba (1), Dots Studio (1), InternScience (1), Ornith AI (1) and Z.AI (1). Strengths differ by task, so check the category leaderboards before choosing by lab.
Explore More
Last updated: November 11, 2026
Know when it’s worth switching models
The model to choose, the cheaper alternative, and the release we would wait on.
Read a sample issueJoin 7,000+ subscribers.
One email each week. Unsubscribe anytime.