Benchmark
ProgramBench
Coding benchmarks compare models on programming tasks. Scores are percentages.
- Models measured
- 41
- Scores recorded
- 41
- Sources
- 2
- IQ weight
- 0.71
Score distribution
- 54.5%
- Mean
- 55.7%
- Median
- 22.2%
- Std dev
- 20.3%
- 10th percentile
- 85.6%
- 90th percentile
- 16%
- Min
This benchmark has a weight of 0.71 for chat models, measured over 41 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.
Leaderboard
| # | Model | |
|---|---|---|
| 1 | Claude Opus 5.5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claimValues from 2 sources
| 91.2% |
| 2 | Muse Wonder Full· ProgramBenchValues from 2 sources
| 88.5% |
| 3 | GPT-6 Astra· ProgramBench | 86.9% |
| 4 | Muse Wonder· ProgramBenchValues from 2 sources
| 86.5% |
| 5 | Claude Fable 5.2· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 85.6% |
| 6 | Claude Fable 5.1· ProgramBench | 82.5% |
| 7 | Claude Haiku 5.5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 82% |
| 8 | Claude Sonnet 5.5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 79.7% |
| 9 | Kimi K3· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 77.8% |
| 10 | Claude Opus 5· ProgramBench | 74.7% |
| 11 | Claude Opus 4.8· ProgramBench | 70.9% |
| 12 | Muse Spark 1.3· ProgramBench | 70.8% |
| 13 | GPT-5.6 Sol· ProgramBench | 69.9% |
| 14 | GPT 5.5· ProgramBench | 69.5% |
| 15 | GLM 5.2· ProgramBenchValues from 2 sources
| 64.6% |
| 16 | Gemini 3.7 Flash· ProgramBench | 61.2% |
| 17 | Muse Spark 1.2· ProgramBench | 57.2% |
| 18 | Gemini 3.6 Flash· ProgramBench | 55.7% |
| 19 | Claude Opus 4.7· ProgramBench | 55.1% |
| 20 | Gemini 3.5 Flash· ProgramBench | 53.6% |
| 21 | Kimi K2.7 Code· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 53.6% |
| 22 | Claude Opus 4.6· ProgramBench | 52.1% |
| 23 | Doubao Seed 2.1 Pro· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 50.3% |
| 24 | Muse Wonder 1.0· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 50% |
| 25 | Doubao Seed 2.1 Turbo· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 49.4% |
| 26 | Claude Sonnet 4.6· ProgramBench | 47.5% |
| 27 | Muse Spark 1.1· ProgramBench | 47% |
| 28 | GPT 5.4· ProgramBench | 37.7% |
| 29 | Gemini 3.1 Pro Preview· ProgramBench | 36.4% |
| 30 | Gemini 3 Flash· ProgramBench | 31.8% |
| 31 | Claude Haiku 4.5· ProgramBench | 30% |
| 32 | DeepSeek V4.1 Flash 1017· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 29.1% |
| 33 | MiMo V2.6 Pro· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 26.5% |
| 34 | MiMo V2.6 Flash· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 26% |
| 35 | LongCat 2.5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 25.5% |
| 36 | DeepSeek V4.1 Flash· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 20.3% |
| 37 | GLM 5.3· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 19% |
| 38 | Hy4 preview· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 17.5% |
| 39 | GPT-5.4 mini· ProgramBench | 16.4% |
| 40 | GPT-5 mini· ProgramBench | 16% |
| 41 | Hy4 Flash· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 16% |
Tokens the model generated per task, as measured by the source. Comparable within this benchmark only.
Data credits: Data by LLM Stats