GPT-6 Astra 在 arXiv 数学基准上小胜 Fable 5.1

Haider. @haider1原推发布于 2026.09.17

在 X 上打开原推

GPT-6 Astra brutally mogged Fable 5.1 on a benchmark testing research-level math problems from recent arXiv papers 88.6% vs. 87.7%, while costing just $3.63/problem vs. $12.77 similar performance at less than one-third the cost

这条是引用着另一条写的：「BrokenArXiv 与 ArXivMath 新基准：Astra 居首」