Leaderboard
Open Weights Rankings
Models with publicly released weights, ranked by their overall Podium Score.
Overall767Coding469Math90Reasoning483Agentic89Knowledge482Multimodal452Coding Agents68Scientific Reasoning96Multilingual12Writing95Chat & Assistants95Research99Tool Calling45Long Context116Speed472Value471Open Weights183
183 models
| # | Model | Podium Score | Arena Elo |
|---|---|---|---|
| 01 | Kimi K3.1 Pro Moonshot AIOpenreasoning | 94.2 | 1490 |
| 02 | Kimi K3.1 Highspeed Moonshot AIOpenreasoning | 91.5 | 1455 |
| 03 | FLUX.2 Dev Black Forest LabsOpen | 90.4 | 1440 |
| 4 | Qwen Audio 3.1 TTS+ AlibabaOpen | 89.4 | 1430 |
| 5 | GLM-4.5 Air Zhipu AIOpenreasoning | 88.9 | 1420 |
| 6 | MiMo v2.6 Flash XiaomiOpen | 88.4 | 1425 |
| 7 | FLUX.2 Klein 9B Black Forest LabsOpen | 88.2 | 1405 |
| 8 | NVIDIA Nemotron 3.5 Super-VL NVIDIAOpen | 87.8 | 1418 |
| 9 | Wan 2.2 A14B AlibabaOpen | 87.4 | 1415 |
| 10 | FLUX.2 Klein 4B Black Forest LabsOpen | 86.5 | 1380 |
| 11 | Kimi K3 Moonshot AIOpenreasoning | 83.3 | 1485 |
| 12 | Qwen3.8 Flash Next AlibabaOpen | 66.8 | 1460 |
| 13 | G9v3-39A5B AI9StarsOpen | 65.0 | — |
| 14 | G9v3-3B AI9StarsOpen | 65.0 | — |
| 15 | HyperNova 60B MultiverseOpen | 65.0 | — |
| 16 | LFM2.5-2.6B Liquid AIOpen | 65.0 | — |
| 17 | LFM2.5-8B-A1B Liquid AIOpen | 65.0 | — |
| 18 | LFM2.5-VL-1.6B Liquid AIOpen | 65.0 | — |
| 19 | LongCat 2.0 LongCatOpen | 65.0 | — |
| 20 | Magistral Small 1.2 MistralOpen | 65.0 | — |
| 21 | Qwen3.8 27B (xhigh) AlibabaOpen | 65.0 | — |
| 22 | Qwen3.8 27B (low) AlibabaOpen | 65.0 | — |
| 23 | Qwen3.8 27B (medium) AlibabaOpen | 65.0 | — |
| 24 | Qwen3.8 27B (Non-reasoning) AlibabaOpenreasoning | 65.0 | — |
| 25 | GLM-5.3 Flash Z.aiOpen | 64.2 | 1445 |
| 26 | GLM-5.2 Z.aiOpen | 59.1 | 1469 |
| 27 | Granite 4.2 30B Instruct IBMOpen | 58.6 | 1380 |
| 28 | Qwen3 8 2 4t A95b AlibabaOpen | 57.7 | — |
| 29 | DeepSeek V4.1 Flash DeepSeekOpenreasoning | 54.5 | 1450 |
| 30 | Glm 5 Z.aiOpen | 52.6 | — |
| 31 | MiMo V2.6 Pro XiaomiOpen | 52.5 | — |
| 32 | Sarvam 105B Sarvam AIOpen | 50.0 | — |
| 33 | Sarvam 30B Sarvam AIOpen | 50.0 | — |
| 34 | DeepSeek V4.1 Flash (Non-reasoning) DeepSeekOpen | 50.0 | — |
| 35 | Granite 4.2 8B Instruct IBMOpen | 49.2 | 1315 |
| 36 | Kimi K2.6 Moonshot AIOpen | 49.0 | 1461 |
| 37 | Sarvam M Reasoning Sarvam AIOpenreasoning | 49.0 | — |
| 38 | MiMo V2.5 Pro XiaomiOpen | 48.0 | 1466 |
| 39 | DeepSeek V4 Pro DeepSeekOpen | 46.0 | 1458 |
| 40 | GLM-5.1 Z.aiOpen | 46.0 | — |
| 41 | DeepSeek V4 Flash 0420 (Non-reasoning) DeepSeekOpen | 46.0 | — |
| 42 | Deepseek V4 Pro 0424 DeepSeekOpen | 45.3 | — |
| 43 | Deepseek V4 Pro 0424 High DeepSeekOpen | 43.7 | — |
| 44 | Deepseek V4 Flash 0420 DeepSeekOpen | 42.1 | — |
| 45 | DeepSeek V4 Flash DeepSeekOpen | 42.0 | 1438 |
| 46 | Granite 4.2 3B Instruct IBMOpen | 42.0 | 1260 |
| 47 | Kimi K2.7 Code Moonshot AIOpen | 41.8 | — |
| 48 | Qwen3 6 Max AlibabaOpen | 41.1 | — |
| 49 | Deepseek V4 Flash 0420 High DeepSeekOpen | 39.0 | — |
| 50 | Deepseek V3 2 Reasoning DeepSeekOpenreasoning | 38.0 | — |
| 51 | Mimo V2 5 0424 XiaomiOpen | 38.0 | — |
| 52 | Qwen3 6 27b AlibabaOpen | 37.7 | — |
| 53 | Glm 5 1 Non Reasoning Z.aiOpenreasoning | 36.3 | — |
| 54 | Gemma 4 31B GoogleOpen | 35.0 | 1451 |
| 55 | Glm 5 2 Non Reasoning Z.aiOpenreasoning | 34.8 | — |
| 56 | Glm 5 Non Reasoning Z.aiOpenreasoning | 33.2 | — |
| 57 | Qwen3 5 397b A17b Non Reasoning AlibabaOpenreasoning | 32.7 | — |
| 58 | Minimax M2 1 MiniMaxOpen | 32.1 | — |
| 59 | Qwen3 6 35b A3b AlibabaOpen | 32.1 | — |
| 60 | Deepseek V4 Pro 0424 Non Reasoning DeepSeekOpenreasoning | 31.9 | — |
| 61 | Mimo V2 Flash Reasoning XiaomiOpenreasoning | 31.9 | — |
| 62 | Qwen3 5 Omni Plus AlibabaOpen | 31.3 | — |
| 63 | Qwen3 6 27b Non Reasoning AlibabaOpenreasoning | 31.3 | — |
| 64 | Deepseek V3 1 Terminus Reasoning DeepSeekOpenreasoning | 31.1 | — |
| 65 | GPT-oss-120B OpenAIOpen | 30.0 | — |
| 66 | Qwen3 5 27b Non Reasoning AlibabaOpenreasoning | 30.0 | — |
| 67 | Deepseek V4 Flash Non Reasoning DeepSeekOpenreasoning | 29.3 | — |
| 68 | Glm 4 6 Reasoning Z.aiOpenreasoning | 29.3 | — |
| 69 | Minimax M2 MiniMaxOpen | 28.9 | — |
| 70 | Mimo V2 5 Pro Non Reasoning XiaomiOpenreasoning | 28.4 | — |
| 71 | Qwen3 5 122b A10b Non Reasoning AlibabaOpenreasoning | 28.2 | — |
| 72 | Glm 4 7 Non Reasoning Z.aiOpenreasoning | 27.1 | — |
| 73 | Gemma 4 26b A4b GoogleOpen | 26.1 | — |
| 74 | Deepseek V3 2 Reasoning 0925 DeepSeekOpenreasoning | 25.9 | — |
| 75 | Qwen3 Max Thinking Preview AlibabaOpenreasoning | 25.5 | — |
| 76 | Qwen3 6 35b A3b Non Reasoning AlibabaOpenreasoning | 24.6 | — |
| 77 | Qwen3 5 35b A3b Non Reasoning AlibabaOpenreasoning | 24.3 | — |
| 78 | Qwen3.5 397B-A17B AlibabaOpen | 23.9 | 1442 |
| 79 | Gemma 4 31b Non Reasoning GoogleOpenreasoning | 22.3 | — |
| 80 | Gemma 4 12b GoogleOpen | 22.2 | — |
| 81 | Qwen3.5 122B-A10B AlibabaOpen | 22.0 | 1417 |
| 82 | GPT-oss-20B OpenAIOpen | 22.0 | — |
| 83 | Qwen3 5 9b AlibabaOpen | 21.8 | — |
| 84 | Deepseek V3 2 0925 DeepSeekOpen | 21.7 | — |
| 85 | Qwen3 Coder Next AlibabaOpen | 21.3 | — |
| 86 | Deepseek V3 1 Reasoning DeepSeekOpenreasoning | 21.0 | — |
| 87 | Qwen3 Vl 235b A22b Reasoning AlibabaOpenreasoning | 20.9 | — |
| 88 | Qwen3 5 9b Non Reasoning AlibabaOpenreasoning | 20.6 | — |
| 89 | Gemma 4 26b A4b Non Reasoning GoogleOpenreasoning | 20.4 | — |
| 90 | Qwen3 5 4b AlibabaOpen | 20.4 | — |
| 91 | Qwen3 235b A22b Instruct 2507 Reasoning AlibabaOpenreasoning | 19.9 | — |
| 92 | Mistral Small 4 MistralOpen | 19.7 | — |
| 93 | Qwen3 Max AlibabaOpen | 19.6 | 1435 |
| 94 | Qwen3 Max Preview AlibabaOpen | 19.4 | — |
| 95 | Qwen3 5 Omni Flash AlibabaOpen | 19.2 | — |
| 96 | Nemotron 3 Ultra 550B NVIDIAOpen | 19.0 | 1426 |
| 97 | Deepseek R1 0120 DeepSeekOpenreasoning | 18.6 | — |
| 98 | Qwen3 235b A22b Instruct 2507 AlibabaOpen | 18.4 | — |
| 99 | Qwen3 Coder 480b A35b Instruct AlibabaOpen | 18.2 | — |
| 100 | Qwen3 Vl 32b Reasoning AlibabaOpenreasoning | 18.1 | — |
Podium Score is a weighted blend of five leaderboards. See Methodology for details.