Best open-weights LLM (October 2026)
As of 21 Oct 2026, the best open-weights LLM on independent benchmarks is Kimi K3 (max) by Moonshot AI (BenchLeader Index 64.9), ahead of MiMo-V2.6-Pro (BenchLeader Index 64.2) and GLM 5.3 (max) (BenchLeader Index 63.6).
Models whose weights you can download and run yourself, ranked on the BenchLeader Index. The licence varies by model; check it on the model page before you ship.
246 models qualify · data as of 21 Oct 2026 · rebuilt once a day (the run starts at 22:03 UTC)
Kimi K3 max#1 here: BenchLeader Index 64.9Best value
DeepSeek V4 Flash maxmost index points per dollar among the top 15 (61.1 at $0.175/M)Fastest of the top 15 DeepSeek V4.1 Flash 1017 max209 output tokens per secondHead to head: Kimi K3 vs MiMo-V2.6-Pro (#1 vs #2) · Kimi K3 vs GLM 5.3 (#1 vs #3) · MiMo-V2.6-Pro vs GLM 5.3 (#2 vs #3)
| # | Model | Index | Price $/M | Speed tok/s |
|---|---|---|---|---|
| 1 | Kimi K3maxMoonshot AI | 64.9 | $6.00 | 42 |
| 2 |  MiMo-V2.6-ProXiaomi | 64.2 | $0.548 | 50 |
| 3 |  GLM 5.3maxZhipu AI | 63.6 | $2.15 | 84 |
| 4 |  Qwen3.8 2.4T A95BAlibaba | 63.2 | $3.00 | 42 |
| 5 | DeepSeek V4.1 Flash 1017maxDeepSeek | 63.1 | $0.525 | 209 |
| 6 | GLM 5.3 FlashZhipu AI | 62.8 | $0.238 | 77 |
| 7 | DeepSeek V4 PromaxDeepSeek | 62.7 | $1.98 | 103 |
| 8 | GLM 5.2maxZhipu AI | 62.4 | $2.15 | 88 |
| 9 |  Muse Spark OpenxhighMeta | 61.2 | $2.00 | 141 |
| 10 | DeepSeek V4 FlashmaxDeepSeek | 61.1 | $0.175 | 145 |
| 11 |  Step 5 PreviewStepFun | 60.6 | $1.43 | 92 |
| 12 | Qwen3.8-Flash-NextAlibaba | 60.3 | $0.230 | 52 |
| 13 | MiMo-V2.6-FlashXiaomi | 59.8 | $0.175 | 53 |
| 14 | Kimi K2.6thinkingMoonshot AI | 59.4 | $1.71 | 72 |
| 15 | GLM-5thinkingZhipu AI | 58.9 | $1.55 | 75 |
Full ranked list (246 models) · Kimi K3 vs MiMo-V2.6-Pro · Methodology · the LLM benchmarks behind this ranking
Best open-weights LLM: questions
- What is the best open-weights LLM in October 2026?
- As of 21 Oct 2026, the best open-weights LLM on independent benchmarks is Kimi K3 (max) by Moonshot AI (BenchLeader Index 64.9), ahead of MiMo-V2.6-Pro (BenchLeader Index 64.2) and GLM 5.3 (max) (BenchLeader Index 63.6). BenchLeader reconciles 229 benchmarks from 43 benchmark publishers and rebuilds the ranking once a day (the run starts at 22:03 UTC).
- Which is the best value?
- DeepSeek V4 Flash (max) gives the most BenchLeader Index points per dollar among the top 15 here: 61.1 at $0.175 per million tokens blended, against $6.00 for Kimi K3.
- How is this ranked?
- Models whose weights you can download and run yourself, ranked on the BenchLeader Index. The licence varies by model; check it on the model page before you ship. Every score comes from an independent publisher and links to its source on the model page; self-reported launch numbers are excluded.
Cite as: BenchLeader, “Best open-weights LLM (October 2026)”, https://www.benchleader.com/best/open-weights-llm, data as of 21 Oct 2026.