Multi-SWE-Bench (c): leaderboard
Metric: Score (%). Source: huggingface.co. 38 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | CodeArts Agent + CodeArts-GLM-5.1 | 54.33 |
| 2 | RepoRepair + Claude-4-Sonnet | 12.61 |
| 3 | MagentLess + Gemini-2.5-Pro | 9.38 |
| 4 | MSWE-agent + Gemini-2.5-Pro | 9.38 |
| 5 | MopenHands + Claude-3.7-Sonnet | 8.59 |
| 6 | MSWE-agent + Claude-3.7-Sonnet | 8.59 |
| 7 | MagentLess + Claude-3.7-Sonnet | 6.25 |
| 8 | MopenHands + Gemini-2.5-Pro | 5.47 |
| 9 | MSWE-agent + Claude-3.5-Sonnet(Oct) | 4.69 |
| 10 | MagentLess + Doubao-1.5-thinking | 4.69 |
Interactive version: theaggregate.ai/benchmark?slug=multi-swe-bench-c · How It Works · Data refreshed daily, snapshot 2026-09-05.