AA LiveCodeBench: leaderboard
Artificial Analysis independent evaluation of LiveCodeBench: contamination-free coding from LeetCode, AtCoder, and CodeForces problems.
Metric: Pass@1 (%). Source: artificialanalysis.ai. Saturation forecast: Estimated already saturated. 343 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Pro (Preview) (High) | 91.75 | #66 (Gemini 3 Pro (Preview)) |
| 2 | DeepSeek V3.2 Speciale | 89.63 | #128 |
| 3 | GPT-5.2 (Medium) | 89.42 | #110 (GPT-5.2) |
| 4 | GLM-4.7 (Reasoning) | 89.42 | #192 (GLM-4.7) |
| 5 | GPT-5.2 (xHigh) | 88.89 | #110 (GPT-5.2) |
| 6 | GPT-OSS-120B (High) | 87.83 | #341 (GPT-OSS-120B) |
| 7 | Claude Opus 4.5 (Thinking) | 87.09 | #83 (Claude Opus 4.5) |
| 8 | GPT-5.1 (High) | 86.77 | #135 (GPT-5.1) |
| 9 | MiMo-V2-Flash (Reasoning) | 86.77 | #372 (MiMo-V2-Flash) |
| 10 | DeepSeek V3.2 (Thinking) | 86.24 | #205 (DeepSeek V3.2) |
| 11 | O4 Mini (High) | 85.93 | #180 (O4 Mini) |
| 12 | Gemini 3 Pro (Preview) (Low) | 85.71 | #66 (Gemini 3 Pro (Preview)) |
| 13 | Kimi K2 (Thinking) | 85.29 | #243 (Kimi K2) |
| 14 | GPT-5.1 Codex (High) | 84.87 | #159 (GPT-5.1 Codex) |
| 15 | GPT-5 (High) | 84.55 | #98 (GPT-5) |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=aa-livecodebench · How It Works · Data refreshed daily, snapshot 2026-10-20.