AI 导读
Epoch AI 在预发布阶段测试了 OpenAI 的 GPT-6 Astra，其 ECI 得分 169，高于此前最高的 163，但仍落在推理时代 ECI 趋势的不确定区间内。Astra 同时在数学、持续学习和游戏解谜基准上刷新最好成绩，在长时程编码基准 MirrorCode 上排名介于 Opus 4.7 与 Fable 5 之间。
正文
GPT-6 Astra has set a new ECI record, with a score of 169. This is a substantial jump from the prior best (163), but is within our uncertainty range for the reasoning-era ECI trend. Astra also set new records on our math, continual learning, and game-puzzles benchmarks. On our long-horizon coding benchmark, MirrorCode, Astra ranks between Opus 4.7 and Fable 5.
OpenAI gave us pre-release access to test Astra. Charts and more details for Astra’s individual benchmark results in the thread.
来源：@EpochAIResearch · x.com