# GPT-6 Astra 数学基准击败 Fable 5.1 且成本仅三分之一

- 来源：Haider. (@haider1)
- 发布时间：2026-09-17 08:50
- AIHOT 分数：38
- AIHOT 链接：https://aihot.news/items/cmu4tql9i07pxrokcfge78n5c
- 原文链接：https://x.com/haider1/status/2100386754423336980

## AI 摘要

GPT-6 Astra 在 BrokenArXiv 与 ArXivMath 基准上以 88.6% 对 87.7% 击败 Fable 5.1，单题成本仅 $3.63，而后者为 $12.77，性能相近但成本不到三分之一。该基准聚焦近一个月 arXiv 上被推翻的猜想，模型在 harness 内执行而非直接调用 API。

## 正文

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

### 引用推文

> Jasper Dekoninck：We are releasing the latest version of BrokenArXiv and ArXivMath! These benchmarks now focus on conjectures that were refuted in the last month on ArXiv, and mo...