Qwen Code Races 5 Models on Your Repo. Its Only Judge Is Told: “Do Not Pick a Winner.”
中文摘要
Qwen Code 评测模型，不指定获胜者。
English Summary
Qwen Code benchmarks models on your repo, with no designated winner.
原文节选
All 11 numbers Agent Arena reports about a candidate come from git, a stopwatch and a token counter. Not one of them runs your code, so I… Continue reading on Towards AI »