[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"tag-humaneval":3},{"tag":4,"articles":10},{"id":5,"name":6,"slug":7,"article_count":8,"description_zh":9,"description_en":9},"b3dfc324-7aa3-4d3e-b73e-7b65d03a67e7","HumanEval","humaneval",0,null,[11],{"id":12,"slug":13,"title":14,"summary":15,"category":16,"image_url":17,"cover_image":17,"language":18,"created_at":19},"5b168b94-465a-4d72-bbb1-e6577625cb1a","why-coding-benchmarks-are-finally-telling-the-truth-zh","為什麼程式碼基準測試終於開始說實話","BenchLM 的程式碼排行榜顯示，真正有用的訊號只剩 LiveCodeBench 與 SWE-bench Pro；HumanEval 已經不適合拿來選模型。","research","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1778670697069-56o7.png","zh","2026-05-13T11:10:25.586869+00:00"]