LiveCodeBench Leaderboard (default window 8/1/2024–5/1/2025, 454 problems)
LiveCodeBench Leaderboard (default window 8/1/2024–5/1/2025, 454 problems) is run by LiveCodeBench. It has ranked 28 models, of which the catalogue holds 26, scoring from 20.2 to 80.2 on Pass@1.
measured by LiveCodeBench · the board itself
| Place | Model | Metric | Score |
|---|---|---|---|
| 1stof 28 | o4 Mini High | Pass@1 | 80.2 |
| 2ndof 28 | o3 | Pass@1 (high) | 75.8 |
| 3rdof 28 | o4-mini | Pass@1 | 74.2 |
| 4thof 28 | Gemini 2.5 Pro Preview 06-05 | Pass@1 | 73.6 |
| 5thof 28 | DeepSeek-R1 (0528) | Pass@1 | 73.1 |
| 6thof 28 | Gemini 2.5 Pro Preview 05-06 | Pass@1 | 71.8 |
| 7thof 28 | EXAONE-4.0-32B | Pass@1 | 70 |
| 8thof 28 | OpenReasoning-Nemotron-32B | Pass@1 | 69.8 |
| 9thof 28 | o3-mini | Pass@1 (high) | 67.4 |
| 10thof 28 | OpenCodeReasoning-Nemotron-1.1-32B | Pass@1 | 66.8 |
| 11thof 28 | Grok 3 Mini | Pass@1 | 66.7 |
| 13thof 28 | Qwen3 235B A22B | Pass@1 | 65.9 |
| 14thof 28 | XBai o4 | Pass@1 (medium) | 65 |
| 15thof 28 | o3-mini | Pass@1 (med) | 63 |
| 16thof 28 | Gemini 2.5 Flash | Pass@1 | 61.9 |
| 18thof 28 | o3-mini | Pass@1 (low) | 57 |
| 19thof 28 | Claude Opus 4 | Pass@1 (thinking) | 56.6 |
| 19thof 28 | Claude Opus 4 | Pass@1 | 56.6 |
| 20thof 28 | Claude Sonnet 4 | Pass@1 (thinking) | 55.9 |
| 20thof 28 | Claude Sonnet 4 | Pass@1 | 55.9 |
| 23rdof 28 | Claude 3.5 Sonnet | Pass@1 | 36.4 |
| 24thof 28 | GPT-4o (2024-08-06) | Pass@1 | 29.5 |
| 25thof 28 | GPT-4 Turbo | Pass@1 | 28.7 |
| 26thof 28 | GPT-4o-mini (2024-07-18) | Pass@1 | 27.5 |
| 27thof 28 | DeepSeek-V3 | Pass@1 | 27.2 |
| 28thof 28 | Claude 3 Haiku | Pass@1 | 20.2 |
Read from the board on 2026-08-25