项目精选
bridgebench
模型在真实工程任务上两两对决,盲审评分,Elo 排名。
01
项目亮点
- 336 个公共任务
- 七领域覆盖
- 盲审三人评委
- 可审计日志
FACTS
来自当前修订的事实
- repository owner
- bridge-mind
- repository name
- bridgebench
- description
- The world's #1 vibe coding benchmark — models go head-to-head on real engineering tasks, judged blind, ranked by Elo.
- primary language
- TypeScript
- license
- MIT
- topics
- ai, benchmark, evaluation, leaderboard, llm, vibe-coding
- source count
- 1