Qwen Code Scores 76.16% on Full SWE-bench Verified Run, Resolving 377 of 500
DennisYu07 · ghdev · 2026-08-21
QwenLM/qwen-code posted end-to-end full benchmark results (v0.21.15, model qwen3.7-plus):
- SWE-bench Verified: all 500 tasks completed, 377 resolved, 118 unresolved, 2 execution errors, 3 infrastructure failures, scoring 76.16% (denominator: 495 valid grader results).
- Followed by Terminal-Bench 2.0 (89 tasks) with verifier-backed results and trajectory writeback.
- Case trajectories are packaged as swe-bench-verified-dsw-eas-full-20260821-r1-trajectories.tar.gz; the full run is available on GitHub Actions.
More from coding & agent
- Ex-DeepMind Eng Lead's AI Calling Agent Fo Cut Hang-Up Rate from 80% to 20-30% — julianweisser · 2026-10-16
- SPACE Open-Sources Its Sandbox Platform for Agentic Code Execution — marktenenholtz · 2026-10-16
- Cognition Launches Devin OnCall to Auto-Investigate Production Alerts — DevinAI · 2026-10-16
- Devin Oncall tops SREGym with 100% end-to-end success, auto-resolving production incidents — tianyin_xu · 2026-10-16
- Running a multi-model UI design arena: Paseo orchestrates four design drafts per model — vista8 · 2026-10-16
- Cortex A/B-tests coding-agent skills and rules by replaying developer corrections before keeping them — Samuele Giampieri · 2026-10-16