LiveCodeBench leaderboard
LiveCodeBench is a contamination-free coding benchmark composed of programming problems collected after model training cut-offs.
72 models ranked, highest score first.
| # | Model | Company | Score |
|---|---|---|---|
| 1 | GPT-5.2 | OpenAI | 88.9% |
| 2 | Claude Opus 4.5 | Anthropic | 87.1% |
| 3 | GPT-5.1 | OpenAI | 86.8% |
| 4 | DeepSeek V3.2 | DeepSeek | 86.2% |
| 5 | GPT-5.3 Codex | OpenAI | 84.2% |
| 6 | GPT-5.2 Codex | OpenAI | 80.4% |
| 7 | Gemini 3.1 Pro | 78.9% | |
| 8 | GPT-5.5 | OpenAI | 74.3% |
| 9 | MiniMax-M3 | MiniMax | 74.0% |
| 10 | MiMo-V2.5-Pro | Xiaomi | 73.3% |
| 11 | Claude Opus 4.7 | Anthropic | 70.3% |
| 12 | Claude Fable 5 | Anthropic | 70.0% |
| 13 | Qwen3.6 Max Preview | Alibaba | 69.7% |
| 14 | Kimi K2.6 | Kimi | 69.7% |
| 15 | Muse Spark | Meta | 69.7% |
| 16 | Qwen3.6 Plus | Alibaba | 69.7% |
| 17 | Gemini 2.5 Flash | 69.5% | |
| 18 | Qwen3.7 Max | Alibaba | 69.0% |
| 19 | Kimi K2 | Moonshot AI | 68.9% |
| 20 | Qwen3.6 27B | Alibaba | 68.7% |
| 21 | Claude Opus 4.8 | Anthropic | 67.7% |
| 22 | DeepSeek V4 Pro | DeepSeek | 66.3% |
| 23 | Qwen3.7 Plus | Alibaba | 65.0% |
| 24 | Ring-2.6-1T | InclusionAI | 64.3% |
| 25 | Grok 4.3 | xAI | 64.3% |
| 26 | Step 3.7 Flash | StepFun | 63.7% |
| 27 | Qwen3.6 35B A3B | Alibaba | 63.7% |
| 28 | DeepSeek V4 Flash | DeepSeek | 63.0% |
| 29 | MiMo-V2.5 | Xiaomi | 62.7% |
| 30 | GLM-5.1 | Z AI | 62.3% |
| 31 | Gemma 4 31B | 62.0% | |
| 32 | Mistral Medium 3.5 | Mistral | 61.0% |
| 33 | GLM 5V Turbo | Z AI | 61.0% |
| 34 | Claude Sonnet 4.6 | Anthropic | 58.7% |
| 35 | Grok 4.20 0309 v2 | xAI | 58.0% |
| 36 | GPT-5 | OpenAI | 55.8% |
| 37 | GPT-5.5 Instant | OpenAI | 55.7% |
| 38 | Gemma 4 26B A4B | 55.7% | |
| 39 | Gemma 4 12B | 55.3% | |
| 40 | JT-35B-Flash | China Mobile | 55.3% |
| 41 | Hy3-preview | Tencent | 54.7% |
| 42 | Step 3.5 Flash 2603 | StepFun | 54.3% |
| 43 | Gemini 3.5 Flash | 53.3% | |
| 44 | EXAONE 4.5 33B | LG AI Research | 49.3% |
| 45 | Mistral Large 3 | Mistral | 46.5% |
| 46 | Qwen3.5 Omni Flash | Alibaba | 44.0% |
| 47 | Claude 3.5 Sonnet | Anthropic | 38.1% |
| 48 | DeepSeek-V3 | DeepSeek | 35.9% |
| 49 | Ling-2.6-1T | InclusionAI | 34.7% |
| 50 | Trinity Large Thinking | Arcee AI | 33.0% |
| 51 | North Mini Code | Cohere | 32.3% |
| 52 | HyperNova 60B 2605 | Multiverse Computing | 31.7% |
| 53 | Gemma 4 E4B | 30.7% | |
| 54 | GPT-4 Turbo | OpenAI | 29.1% |
| 55 | Claude 3 Opus | Anthropic | 27.9% |
| 56 | Solar Pro 3 | Upstage | 27.0% |
| 57 | Grok-2 | xAI | 26.7% |
| 58 | Ling 2.6 Flash | InclusionAI | 25.0% |
| 59 | Gemini 1.5 Pro | 24.4% | |
| 60 | Gemini 2.0 Flash | 21.0% | |
| 61 | Claude 2.1 | Anthropic | 19.5% |
| 62 | Granite 4.1 30B | IBM | 18.7% |
| 63 | Mistral Large | Mistral | 17.8% |
| 64 | Claude 3 Sonnet | Anthropic | 17.5% |
| 65 | Claude 3 Haiku | Anthropic | 15.4% |
| 66 | Gemma 4 E2B | 15.0% | |
| 67 | Granite 4.1 8B | IBM | 12.0% |
| 68 | JT-MINI | China Mobile | 11.7% |
| 69 | MiniCPM-V 4.6 1.3B | OpenBMB | 6.3% |
| 70 | MiniCPM5-1B | OpenBMB | 3.7% |
| 71 | Granite 4.1 3B | IBM | 3.0% |
| 72 | LFM2.5-8B-A1B | Liquid AI | 0.0% |