WHAT NEW IN X11.07 SAT

X 上的 AI 合集档案 · 23 个合集 · 70,791 条

gpt-6POST

公开模型之外:OpenAI 与 Anthropic 内部模型领先多少

作者比较 Opus 5.5、Fable 5.1 与 GPT-6 Astra,并推测两家内部模型大约领先公开版两个月。

Haider.
@haider12026.09.24收录原推发布于 2026.09.24
Haider. @haider1原推发布于 2026.09.24
在 X 上打开原推

opus 5.5, fable 5.1, and gpt-6 astra — and we're only comparing the public models here both openai and anthropic have more advanced internal models. oai showed internal-model performance graphs around the navier-stokes work, which could give us a rough idea of how far ahead they are internally probably around 2 months ahead?

x.com/haider1/status/2102793866067972110采集于 2026.09.24 10:16