SWE-bench Verified
Code · 2024-08-13
SWE-bench Verified is a human-validated subset of 500 SWE-bench tasks drawn from real GitHub issues in Python repositories. It measures the percentage of issues resolved by a model-generated patch that passes the repository tests.
Top models (higher is better)
| Model | Score |
|---|---|
| Claude Opus 5.5 | 98.2 |
| Claude Opus 5 | 97.0 |
| Claude Sonnet 5.5 | 96.7 |
| Claude Fable 5.1 | 96.6 |
| DeepSeek V4 Pro 0813 | 96.4 |
| GPT-5.6 Sol | 96.2 |
| Grok 4.6 | 95.6 |
| Claude Fable 5 | 95.5 |
| GLM-5.3 | 95.4 |
| GPT-5.6 Terra | 95.4 |
| Claude Mythos Preview | 93.9 |
| Kimi K3 | 93.4 |
| GPT-5.6 Luna | 93.0 |
| Ember-1 | 92.2 |
| GLM-5.3 Flash | 92.0 |
| Claude Haiku 5.5 | 88.9 |
| DeepSeek-V4-Flash-0731 | 88.8 |
| Opus 4.8 | 88.6 |
| GPT-6.1 Sol | 87.3 |
| Grok 4.5 | 86.6 |
| Muse Spark 1.2 | 86.6 |
| Qwen3.8 27B | 86.0 |
| Qwen3.8-Max | 85.6 |
| Macaron-V1-Venti | 85.6 |
| DeepSeek V4.1 Flash | 82.9 |
| GLM-5.2 | 82.8 |
| GPT-5.5 | 82.6 |
| Inkling-Small | 82.2 |
| Opus 4.7 | 82.0 |
| Muse Spark 1.1 | 82.0 |
| Beam 1.0 501B A23B | 81.6 |
| Gemini 3.6 Flash | 81.1 |
| Gemini 3.7 Flash | 81.0 |
| Opus 4.5 | 80.9 |
| Opus 4.6 | 80.8 |
| Gemini 3.8 Flash | 80.6 |
| Gemini 3.1 Pro Preview | 80.6 |
| GPT-6 Sol | 80.3 |
| GPT-5.2 | 80.0 |
| Sonnet 4.6 | 79.6 |