Unverified50% confidenceFactExact time
SWE-bench 的评测任务来自 12 个主流 Python 开源仓库，包括 Django、Flask、scikit-learn 等
1
Sources
50%
Confidence
Long-term
Relevance
7/12/2026
First Seen
Sources
Grok 4.5 Hands-On: A Faster, Cheaper Coding Dark Horse — Can It Replace Opus and GPT?
youtubeForrestKnight7/10/2026
Related Claims
UnverifiedThe SWE-Bench test set is sourced from real bug reports from mainstream Python open-source projects such as Django, Flask, NumPy, and Scikit-learn.84% similarPartially VerifiedProjects collected by SWE-bench include well-known Python open-source projects such as Django, scikit-learn, and sympy82% similarUnverified该工具基于Python和Flask构建69% similarPartially VerifiedThe SWE-Bench test set contains 2,294 real bug-fix tasks from 12 mainstream Python open-source projects.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/488745API
curl https://kongchang.com/api/v1/knowledge/claims/488745MCP
get_claim(id=488745)