OpenMOSS/SWE-bench-Science

A benchmark for testing whether coding agents can resolve engineering tasks in scientific software

Stars: 80
Forks: 4
Open beginner issues: 0
Indexed issues: 2

Dominant language: Python
License: MIT
Last GitHub push: Sep 6, 2026
Latest indexed: Sep 19, 2026
Contributing guide: No contributing guide
Code of conduct: No code of conduct
Beginner labels: No beginner labels indexed
PR merge metrics: No merged PRs in 30d

Beginner friendly Everything

Open All

2 open issues indexed

Official DeepSeek v4 Pro Max Trajectories: 3 of 50 Passes Contaminated by Proxy-Bypass Network Exfiltration, Screened Score 47/119 Open

Difficulty 4/5 3-5 days Newbie friendliness 42/100

OpenMOSS/SWE-bench-Science#2 · 1 comment · Opened Sep 7, 2026
Plan to open-source data generation code & clarification around "expert-exploratory" Open

Difficulty 5/5 Over a week Newbie friendliness 20/100

OpenMOSS/SWE-bench-Science#1 · Opened Aug 25, 2026

A short digest of beginner-friendly GitHub issues.