DeepSWE: leaderboard
DataCurve benchmark where frontier coding agents work through original, long-horizon software engineering tasks in realistic repositories.
Metric: Pass@1 (%; DeepSWE v1.1 scored attempts). Source: deepswe.datacurve.ai. Source: deepswe.datacurve.ai. Saturation forecast: Around January 2027. 70 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-6 Astra (xHigh) | 74.1 | #6 (GPT-6 Astra) |
| 2 | Gemini 3.8 Flash (High) | 73.8 | #21 (Gemini 3.8 Flash) |
| 3 | Claude Opus 5 (Max) | 73.6 | #12 (Claude Opus 5) |
| 4 | GPT-6 Astra (Max) | 73.2 | #6 (GPT-6 Astra) |
| 5 | GPT-6 Astra (High) | 73.2 | #6 (GPT-6 Astra) |
| 6 | Claude Opus 5 (xHigh) | 73.2 | #12 (Claude Opus 5) |
| 7 | Claude Opus 5 (High) | 72.8 | #12 (Claude Opus 5) |
| 8 | GPT-6 Astra (Medium) | 72.8 | #6 (GPT-6 Astra) |
| 9 | GPT-5.6 Sol (Max) | 72.7 | #19 (GPT-5.6 Sol) |
| 10 | Gemini 3.8 Flash (Medium) | 71 | #21 (Gemini 3.8 Flash) |
| 11 | GPT-5.6 Sol (xHigh) | 70.7 | #19 (GPT-5.6 Sol) |
| 12 | Claude Fable 5 (xHigh) | 69.9 | #16 (Claude Fable 5) |
| 13 | Claude Fable 5 (Max) | 69.7 | #16 (Claude Fable 5) |
| 14 | GPT-5.6 Terra (Max) | 69.6 | #41 (GPT-5.6 Terra) |
| 15 | GPT-5.6 Sol (High) | 69.4 | #19 (GPT-5.6 Sol) |
No result here: #3 Claude Opus 5.5, #7 GPT-6.1 Sol, #8 Gemini 4 Argon.
Interactive version: theaggregate.ai/benchmark?slug=deepswe · How It Works · Data refreshed daily, snapshot 2026-10-20.