有人抱怨 Codex 现在打不过 Opus 5.5

Irushi @Im_IrushiK原推发布于 2026.09.23

在 X 上打开原推

Codex is in a really rough spot atm. AGREE?? For a long time, OpenAI had the better coding models and better usage limits, which made Codex an easy choice for heavy coding. Now Opus 5.5 feels like it has flipped that. It’s faster, extremely strong at coding, and the limits feel significantly more generous. Meanwhile, Astra is capable but painfully slow, Sol 6 hasn’t impressed me nearly as much, and Codex usage disappears ridiculously fast with heavy use. This might be the weakest Codex has felt relative to the competition in a long time. OpenAI really needs a strong Astra/Sol 6.1 update, but just as importantly, much better usage limits.