Benchmark
SWE-Marathon
Coding benchmarks compare models on programming tasks. Scores are percentages. Higher is better.
- Models measured
- 13
- Scores recorded
- 18
- Sources
- 2
- IQ weight
- 0.77
Score distribution
- 35.8%
- Mean
- 42%
- Median
- 13.7%
- Std dev
- 16%
- 10th percentile
- 53.1%
- 90th percentile
- 13%
- Min
This benchmark has a weight of 0.77 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 | Muse Wonder Full· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 57.5% |
| 2 | Muse Wonder· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 53.1% |
| 3 | Muse Wonder 1.0· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 53.1% |
| 4 | Kimi K3.1 Pro· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 51.7% |
| 5 | Grok 4.7· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 46% |
| 6 | GLM 5.3· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 42.5% |
| 7 | GLM 5.3v1.1 · Claude Code 2.1.207 · max · temp 1.0 · top_p .95 · 1M · 128K output· Reported by Z.ai· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 42.5% |
| 8 | Kimi K3· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 42% |
| 9 | Kimi K3H20-calibrated pre-v1.1 task snapshot · max effort · temp 1.0 · top_p 1.0· Reported by Moonshot AI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 42% |
| 10 | Hy4 preview· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 31.9% |
| 11 | Hy4 previewofficial model-card appendix· Reported by Tencent Hy· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 31.9% |
| 12 | Grok 4.5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 29% |
| 13 | Grok 4.5resolution rate · pass@1· Reported by SpaceXAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 29% |
| 14 | Claude Opus 4.8max· Reported by SpaceXAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 26% |
| 15 | Claude Fable 5max· Reported by SpaceXAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 24% |
| 16 | Claude Opus 4.7max· Reported by SpaceXAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 16% |
| 17 | GLM 5.2· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 13% |
| 18 | GLM 5.21M context · max effort · 128K output· Reported by Z.ai· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 13% |
Data credits: 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. · Data by LLM Stats