SWE-bench Multilingual: leaderboard
Multilingual SWE-bench extension with software issue-resolution tasks spanning multiple programming languages and repository ecosystems.
Metric: Resolved (%). Source: www-cdn.anthropic.com. Status: years away from saturation. 2 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.8 | 84.4 |
| 2 | Claude Opus 4.7 | 80.5 |
Interactive version: theaggregate.ai/benchmark?slug=swe-bench-multilingual · How It Works · Data refreshed daily, snapshot 2026-09-05.