Benchmark
MBPP
Coding benchmarks compare models on programming tasks. Scores are percentages.
- Models measured
- 65
- Scores recorded
- 65
- Sources
- 2
- IQ weight
- 1.60
Score distribution
- 72.3%
- Mean
- 74.4%
- Median
- 15.1%
- Std dev
- 52.4%
- 10th percentile
- 89.7%
- 90th percentile
- 29.8%
- Min
This benchmark has a weight of 1.60 for chat models, measured over 64 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 | O1 Preview· EvalPlus Leaderboard data | 95.5% |
| 2 | O1 Mini· EvalPlus Leaderboard data | 93.1% |
| 3 | Sarvam-30B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 92.7% |
| 4 | Llama 3.3 Nemotron Super 49B v1· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 91.3% |
| 5 | Qwen 2.5 Coder 32B Instruct· EvalPlus Leaderboard dataValues from 2 sources
| 90.5% |
| 6 | gemini-1.5-pro-002· EvalPlus Leaderboard data | 89.7% |
| 7 | Claude 3 Opus· EvalPlus Leaderboard data | 89.4% |
| 8 | DeepSeek Coder V2 Instruct· EvalPlus Leaderboard data | 89.4% |
| 9 | Qwen2.5 72B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 88.2% |
| 10 | DeepSeek V2.5· EvalPlus Leaderboard data | 87.6% |
| 11 | DeepSeek V3· EvalPlus Leaderboard data | 87.6% |
| 12 | GPT 4o· EvalPlus Leaderboard data | 87.6% |
| 13 | Grok Beta· EvalPlus Leaderboard data | 86% |
| 14 | GPT-4 Turbo· EvalPlus Leaderboard data | 85.7% |
| 15 | gemini-1.5-flash-002· EvalPlus Leaderboard data | 84.7% |
| 16 | Qwen2.5 32B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 84% |
| 17 | Claude 3 Sonnet· EvalPlus Leaderboard data | 83.6% |
| 18 | Qwen2.5 Coder 7B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 83.5% |
| 19 | GPT-3.5 Turbo· EvalPlus Leaderboard data | 82.5% |
| 20 | Llama 3 70B Instruct· EvalPlus Leaderboard data | 82.3% |
| 21 | Qwen 2.5 14B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 82% |
| 22 | Qwen 3 235B A22B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 81.4% |
| 23 | phi-3.5-moe-instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 80.8% |
| 24 | DeepSeek Coder 33B Instruct· EvalPlus Leaderboard data | 80.4% |
| 25 | Claude 3 Haiku· EvalPlus Leaderboard data | 80.2% |
| 26 | Qwen2 72B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 80.2% |
| 27 | Qwen2.5 7B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 79.2% |
| 28 | Llama 4 Maverick· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 77.6% |
| 29 | deepseek-coder-7b-instruct-v1p5· EvalPlus Leaderboard data | 75.2% |
| 30 | deepseek-coder-6.7b-instruct· EvalPlus Leaderboard data | 74.9% |
| 31 | Mistral Small 3.1 24b Instruct 2503· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 74.7% |
| 32 | Gemma 3 27B IT· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 74.4% |
| 33 | Command R+· EvalPlus Leaderboard data | 74.3% |
| 34 | Mixtral 8x22B Instruct v0.1· EvalPlus Leaderboard data | 73.8% |
| 35 | Qwen2.5-Omni 7B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 73.2% |
| 36 | Gemma 3 12B IT· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 73% |
| 37 | Mistral Large· EvalPlus Leaderboard data | 72.8% |
| 38 | Qwen1.5-72B-Chat· EvalPlus Leaderboard data | 72.5% |
| 39 | dolphin-2p6-mixtral-8x7b· EvalPlus Leaderboard data | 70.6% |
| 40 | code-llama-34b· EvalPlus Leaderboard data | 69.3% |
| 41 | Llama 3.1 8B Instruct· EvalPlus Leaderboard data | 68.3% |
| 42 | Llama 4 Scout· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 67.8% |
| 43 | DBRX-instruct· EvalPlus Leaderboard data | 67.2% |
| 44 | Qwen2 7B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 67.2% |
| 45 | Phi-3 Mini 4K Instruct· EvalPlus Leaderboard data | 65.9% |
| 46 | codegemma-7b· EvalPlus Leaderboard data | 65.1% |
| 47 | Llama 3 8B Instruct· EvalPlus Leaderboard data | 64.6% |
| 48 | Gemma 3n E4B IT· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 63.6% |
| 49 | code-llama-13b· EvalPlus Leaderboard data | 63.5% |
| 50 | Gemma 3 4B IT· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 63.2% |
Showing 1–50 of 65 results.
Data credits: EvalPlus Leaderboard and EvalPerf (evalplus.github.io), Apache-2.0; EvalPerf win-rate fractions converted to percentages (licence) · Data by LLM Stats