Benchmark
LiveCodeBench
Coding benchmarks compare models on programming tasks. Scores are percentages. Higher is better.
- Models measured
- 81
- Scores recorded
- 98
- Sources
- 3
- IQ weight
- 1.24
Score distribution
- 61.3%
- Mean
- 65.7%
- Median
- 22.1%
- Std dev
- 29.7%
- 10th percentile
- 84.7%
- 90th percentile
- 1.9%
- Min
This benchmark has a weight of 1.24 for chat models, measured over 78 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.
Leaderboard
| # | Model | Score |
|---|---|---|
| 1 | Fugu UltraFugu official headline evaluation· Reported by Sakana AI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 93.2% |
| 2 | DeepSeek V4 FlashDeepSeek V4 official model card · Think Max· Reported by DeepSeek· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 91.6% |
| 3 | Qwen3.7 MaxQwen3.7-Max release · recommended reasoning-effort=xhigh system prompt· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 91.6% |
| 4 | Qwen3.7 PlusQwen3.7-Plus release · recommended reasoning-effort=xhigh system prompt· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 89.6% |
| 5 | DeepSeek V3.2 SpecialeSpeciale high-compute reasoning · temperature 1.0 · 128K context · pass@1-CoT· Reported by DeepSeek· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 88.7% |
| 6 | DeepSeek V4 Flash· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 88.4% |
| 7 | DeepSeek V4 FlashDeepSeek V4 official model card · Think High· Reported by DeepSeek· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 88.4% |
| 8 | Beam 1.0 501B A23B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 88.3% |
| 9 | Solar Pro 4· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 87.8% |
| 10 | o4 Minioverall, high· LiveCodeBench Leaderboard data | 87.3% |
| 11 | OpenAI o3overall, high· LiveCodeBench Leaderboard data | 84.7% |
| 12 | o4 Minioverall, medium· LiveCodeBench Leaderboard data | 84.5% |
| 13 | DeepSeek R1overall· LiveCodeBench Leaderboard data | 84.4% |
| 14 | Gemini 2.5 Prooverall· LiveCodeBench Leaderboard data | 84.3% |
| 15 | DeepSeek Reasoner· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 83.3% |
| 16 | DeepSeek V3.2· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 83.3% |
| 17 | DeepSeek V3.2Thinking mode · temperature 1.0 · 128K context · pass@1-CoT· Reported by DeepSeek· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 83.3% |
| 18 | MiMo V2 Flashtechnical report Table 7 · student after MOPD · benchmark version unspecified· Reported by Xiaomi MiMo· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 83.2% |
| 19 | MiniMax M2· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. | 83% |
| 20 | LongCat-Flash-Thinking· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 82.8% |
| 21 | Nemotron 3 Super 120B A12B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 81.2% |
| 22 | Nemotron 3 Super 120B A12BNeMo Evaluator / mostly NeMo Skills · v5 · 2024-07↔2024-12· Reported by NVIDIA· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 81.2% |
| 23 | Grok 3 Mini· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 80.4% |
| 24 | Qwen 3 235B A22Boverall· LiveCodeBench Leaderboard data | 80.4% |
| 25 | Grok 4 Fast· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 80% |
| 26 | Grok 3· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 79.4% |
| 27 | Grok 4· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 79% |
| 28 | Grok 3 Minioverall, high· LiveCodeBench Leaderboard data | 78.1% |
| 29 | MiniMax M2.1· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 78% |
| 30 | o3 Minioverall, high· LiveCodeBench Leaderboard data | 77.7% |
| 31 | o4 Minioverall, low· LiveCodeBench Leaderboard data | 77.4% |
| 32 | Gemini 2.5 Flashoverall· LiveCodeBench Leaderboard data | 76.2% |
| 33 | o3 Minioverall, Med· LiveCodeBench Leaderboard data | 75.4% |
| 34 | DeepSeek V3.2 Exp· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 74.1% |
| 35 | DeepSeek R1· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 73.3% |
| 36 | GLM 4.5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 72.9% |
| 37 | Gemini 3.1 Flash LiteGemini 3.1 Flash-Lite Model Card · high thinking · default sampling · single attempt/pass@1 · 175 problems · 2025-01-01 through 2025-05-01· Reported by Google DeepMind· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 72% |
| 38 | Nemotron Nano 9b V2· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 71.1% |
| 39 | Nova 2 Lite· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 71% |
| 40 | GLM 4.5 Air· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 70.7% |
| 41 | Qwen 3 235B A22B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 70.7% |
| 42 | o3 Minioverall, low· LiveCodeBench Leaderboard data | 70.6% |
| 43 | Claude Opus 4overall, thinking· LiveCodeBench Leaderboard data | 70.4% |
| 44 | Gemini 2.5 Pro· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 69% |
| 45 | Claude Sonnet 4overall, thinking· LiveCodeBench Leaderboard data | 68.5% |
| 46 | Mercury 2· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 67% |
| 47 | Llama 3.1 Nemotron Ultra 253B v1· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 66.3% |
| 48 | Qwen3 32B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 65.7% |
| 49 | MiniMax M1 80K· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. | 65% |
| 50 | Ministral 14Blatest· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 64.6% |
Showing 1–50 of 98 results.
Data credits: Data by LLM Stats · BenchBoard by SYuan03 (github.com/SYuan03/benchboard). MIT covers its code and data organization; no results licence is inferred. Numeric relays retain the original stating-source links.