SWE-PolyBench Verified (Python): leaderboard
SWE-PolyBench Verified Python slice evaluates agents on resolving Python GitHub issues from the verified set.
Metric: Score (%). Source: huggingface.co. 5 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | HMigBot | 61.95 | |
| 2 | Atlassian Rovo Dev (2025-12-08) | 54.87 | |
| 3 | Prometheus-v1.2 + GPT-5 | 36.28 | |
| 4 | Kodah | 32.74 | |
| 5 | iSWE-Agent | 0 |
Interactive version: theaggregate.ai/benchmark?slug=swe-polybench-verified-python · How It Works · Data refreshed daily, snapshot 2026-10-20.