DeepSWE — leaderboard
DataCurve benchmark where frontier coding agents work through original, long-horizon software engineering tasks in realistic repositories.
Metric: Pass@1 (%). Source: deepswe.datacurve.ai. Status: saturation imminent. 21 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.6 Sol | 72.7 |
| 2 | GPT-5.6 Sol (xHigh) | 70.7 |
| 3 | Claude Fable 5 | 69.9 |
| 4 | Kimi K3 | 68.5 |
| 5 | GPT-5.5 (xHigh) | 67 |
| 6 | GPT-5.5 | 64.4 |
| 7 | Claude Opus 4.8 | 59 |
| 8 | Claude Sonnet 5 | 53.8 |
| 9 | Grok 4.5 | 53.8 |
| 10 | Muse Spark 1.1 | 53.3 |
| 11 | GPT-5.4 (xHigh) | 51.8 |
| 12 | Gemini 3 Flash | 48.6 |
| 13 | GLM-5.2 | 43.8 |
| 14 | Gemini 3.5 Flash | 37.4 |
| 15 | Kimi K2.7 Code | 30.5 |
Interactive version: theaggregate.ai/benchmark?slug=deepswe · How the rankings work · Data refreshed daily, snapshot 2026-07-22.