Peter Steinberger(@steipete)
RT @Kaju_Nut: "Our initial evaluation suggested that GPT-6-Astra's score was in the late eighties followed closely by Opus 5.5. But, as we…
6.4内容质量
TL;DR · AI 摘要
Peter Steinberger 🦞 on X: "RT @Kaju Nut: "Our initial evaluation suggested that GPT-6-Astra's score was in the late eigh...
核心要点
- 主题聚焦：RT @Kaju Nut: "Our initial evaluation suggested
- 来源：Peter Steinberger(@steipete)，建议结合原文判断细节。
- AI 分析暂不可用，本条为保底评分与摘要。
#AI
打开原文Peter Steinberger 🦞 on X: "RT @Kaju_Nut: "Our initial evaluation suggested that GPT-6-Astra's score was in the late eighties followed closely by Opus 5.5. But, as we…" / X
Peter Steinberger 🦞 转发了
Nirmalya Kajuri
@Kaju_Nut
4h
"我们的初步评估显示，GPT-6-Astra的得分在八十多分，随后是Opus 5.5。但当我们开始调查模型可能犯的错误时，这些模型，尤其是Astra，坚持认为自己是正确的，而问题出在
显示更多
Pratik Rath
@PratikRath15
14h
arxiv.org/abs/2610.16733
20
172
1.9K
95K