Primary sourcerohanpaul_ai
View sourceEvent ·Multi-source
清华与 Moonshot AI 提出 TCSExam 基准:GPT-6 Astra 定理证明率达 70.0%
清华大学与 Moonshot AI 发布 TCSExam 基准,包含 100 道专家审核任务,覆盖 Lean 定理证明、求解器工程、纪录构造和 30 个理论计算机科学开放猜想四个方向,所有提交由 Lean 内核或任务检查器验证。在无提示引理的条件下,GPT-6 Astra 定理证明成功率为 70.0%,Kimi K3.1 Pro 为 38.3%,其余 6 个智能体均不超过 11.7%。但在 30 个开放猜想上,GPT-6 Astra 一个都未能解决,仅有的几次被接受的提交改编自 OpenAI 已公开的反例与证明。
Current read
清华和 Moonshot AI 搞了个 TCSExam 测试集,GPT-6 Astra 证定理能拿 70.0%,但 30 个开放猜想全挂,差距很直观。
11 sources69° AI heatLast updated 2026/11/8 18:45:52
Evidence
11 sourcesRelated source 1lmarena.ai
View sourceRelated source 2Decoder
View sourceRelated source 3AI Will
View sourceRelated source 4OpenAI
View sourceRelated source 5Akshay Kothari
View sourceRelated source 6shao__meng
View sourceRelated source 7Tech in Asia
View sourceRelated source 8Andrew Curran
View sourceRelated source 9@koltregaskes
View sourceRelated source 10TestingCatalog
View sourceConflict check
Dedup data only shows that these sources discuss the same event, not that they agree. There is no structured supporting or refuting evidence yet, so no conflict is inferred automatically.