Multi-SWE-Bench (typescript): leaderboard
Multi-SWE-Bench TypeScript slice evaluates agents on resolving TypeScript GitHub issues.
Metric: Score (%). Source: huggingface.co. 37 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | RepoRepair + Claude-3.5-Sonnet(Oct) | 27.23 | |
| 2 | MopenHands + Gemini-2.5-Pro | 22.32 | |
| 3 | MopenHands + Claude-3.5-Sonnet(Oct) | 11.61 | |
| 4 | MagentLess + Gemini-2.5-Pro | 11.61 | |
| 5 | MSWE-agent + Claude-3.7-Sonnet | 11.16 | |
| 6 | MagentLess + Llama-4-Maverick | 9.38 | |
| 7 | MSWE-agent + Gemini-2.5-Pro | 8.93 | |
| 8 | MSWE-agent + Claude-3.5-Sonnet(Oct) | 8.04 | |
| 9 | MopenHands + Llama-4-Maverick | 8.04 | |
| 10 | MagentLess + Doubao-1.5-thinking | 7.59 |
Interactive version: theaggregate.ai/benchmark?slug=multi-swe-bench-typescript · How It Works · Data refreshed daily, snapshot 2026-10-20.