AA LiveCodeBench: leaderboard
Artificial Analysis independent evaluation of LiveCodeBench: contamination-free coding from LeetCode, AtCoder, and CodeForces problems.
Metric: Pass@1 (%). Source: artificialanalysis.ai. Status: saturation imminent. 181 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Pro (Preview) (High) | 91.75 |
| 2 | GLM-4.7 (Reasoning) | 89.42 |
| 3 | GPT-OSS-120B (High) | 87.83 |
| 4 | Claude Opus 4.5 (Thinking) | 87.09 |
| 5 | GPT-5.1 (High) | 86.77 |
| 6 | MiMo-V2-Flash (Reasoning) | 86.77 |
| 7 | Kimi K2 (Thinking) | 85.29 |
| 8 | GPT-5 (High) | 84.55 |
| 9 | GPT-5 Codex (High) | 84.02 |
| 10 | GPT-5 Mini (High) | 83.81 |
| 11 | Grok 4 Fast (Reasoning) | 83.17 |
| 12 | MiniMax-M2 | 82.65 |
| 13 | Grok 4.1 Fast (Reasoning) | 82.22 |
| 14 | Grok 4 | 81.9 |
| 15 | MiniMax-M2.1 | 80.95 |
Interactive version: theaggregate.ai/benchmark?slug=aa-livecodebench · How It Works · Data refreshed daily, snapshot 2026-09-05.