Benchmark
JudgeBench: Coding Score
Coding benchmarks compare models on programming tasks. Scores are percentages. Higher is better.
- Models measured
- 13
- Scores recorded
- 23
- Sources
- 1
- IQ weight
- 1.14
Score distribution
- 50.6%
- Mean
- 45.2%
- Median
- 26.6%
- Std dev
- 21.4%
- 10th percentile
- 92.9%
- 90th percentile
- 9.7%
- Min
This benchmark has a weight of 1.14 for chat models, measured over 13 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 | o3 MiniGPT-4o Dataset, Prompted Judge, Arena-Hard (o3-mini-2025-01-31 (high)), Coding Score· JudgeBench data docs | 100% |
| 2 | DeepSeek R1GPT-4o Dataset, Prompted Judge, Arena-Hard (DeepSeek-R1-250120), Coding Score· JudgeBench data docs | 92.9% |
| 3 | o3 MiniGPT-4o Dataset, Prompted Judge, Arena-Hard (o3-mini-2025-01-31 (medium)), Coding Score· JudgeBench data docs | 92.9% |
| 4 | O1 PreviewGPT-4o Dataset, Prompted Judge, Arena-Hard (o1-preview-2024-09-12), Coding Score· JudgeBench data docs | 85.7% |
| 5 | o3 MiniGPT-4o Dataset, Prompted Judge, Arena-Hard (o3-mini-2025-01-31 (low)), Coding Score· JudgeBench data docs | 83.3% |
| 6 | O1 MiniGPT-4o Dataset, Prompted Judge, Arena-Hard (o1-mini-2024-09-12), Coding Score· JudgeBench data docs | 78.6% |
| 7 | Claude 3.5 SonnetGPT-4o Dataset, Prompted Judge, Arena-Hard (claude-3-5-sonnet-20240620), Coding Score· JudgeBench data docs | 64.3% |
| 8 | GPT 4oGPT-4o Dataset, Prompted Judge, Vanilla (gpt-4o-2024-05-13), Coding Score· JudgeBench data docs | 61.9% |
| 9 | GPT 4oGPT-4o Dataset, Prompted Judge, Arena-Hard (gpt-4o-2024-05-13), Coding Score· JudgeBench data docs | 59.5% |
| 10 | Llama 3.1 70B InstructGPT-4o Dataset, Prompted Judge, Arena-Hard (Llama-3.1-70B-Instruct), Coding Score· JudgeBench data docs | 52.4% |
| 11 | Llama 3.1 405B InstructGPT-4o Dataset, Prompted Judge, Arena-Hard (Llama-3.1-405B-Instruct), Coding Score· JudgeBench data docs | 50% |
| 12 | GPT-4o miniGPT-4o Dataset, Prompted Judge, Arena-Hard (gpt-4o-mini-2024-07-18), Coding Score· JudgeBench data docs | 45.2% |
| 13 | GPT 4oClaude-3.5-Sonnet Dataset, Prompted Judge, Arena-Hard (gpt-4o-2024-05-13), Coding Score· JudgeBench data docs | 38.7% |
| 14 | Llama 3.1 8B InstructGPT-4o Dataset, Prompted Judge, Arena-Hard (Llama-3.1-8B-Instruct), Coding Score· JudgeBench data docs | 33.3% |
| 15 | Claude 3.5 SonnetClaude-3.5-Sonnet Dataset, Prompted Judge, Arena-Hard (claude-3-5-sonnet-20240620), Coding Score· JudgeBench data docs | 32.3% |
| 16 | GPT 4oGPT-4o Dataset, Multi-Agent Judge, ChatEval (gpt-4o-2024-05-13), Coding Score· JudgeBench data docs | 31% |
| 17 | Llama 3.1 8B InstructClaude-3.5-Sonnet Dataset, Prompted Judge, Arena-Hard (Llama-3.1-8B-Instruct), Coding Score· JudgeBench data docs | 29% |
| 18 | gemini-1.5-pro-001GPT-4o Dataset, Prompted Judge, VertexAI Evaluation (gemini-1.5-pro-001), Coding Score· JudgeBench data docs | 28.6% |
| 19 | gemini-1.5-pro-001GPT-4o Dataset, Prompted Judge, Arena-Hard (gemini-1.5-pro-001), Coding Score· JudgeBench data docs | 26.2% |
| 20 | Llama 3.1 70B InstructClaude-3.5-Sonnet Dataset, Prompted Judge, Arena-Hard (Llama-3.1-70B-Instruct), Coding Score· JudgeBench data docs | 25.8% |
| 21 | Claude 3 HaikuGPT-4o Dataset, Prompted Judge, Arena-Hard (claude-3-haiku-20240307), Coding Score· JudgeBench data docs | 21.4% |
| 22 | gemini-1.5-flash-001GPT-4o Dataset, Prompted Judge, Arena-Hard (gemini-1.5-flash-001), Coding Score· JudgeBench data docs | 21.4% |
| 23 | Claude 3 HaikuClaude-3.5-Sonnet Dataset, Prompted Judge, Arena-Hard (claude-3-haiku-20240307), Coding Score· JudgeBench data docs | 9.7% |
Data credits: JudgeBench, Sijun Tan et al., ScalerLab