用户实测：大重构和长上下文 Gemini 4 Argon 反超 Claude

Se Hozaifa @SeHozaifa原推发布于 2026.10.19

在 X 上打开原推

Lot of people said this one was rage bait, lol its not. Claude is still the pinnacle for a lot of stuff, Opus 5.5 still tops Terminal-Bench no doubt. Just my real experience using both. Claude was my fav too but lately on big refactors and long context Gemini felt ahead. Genuinely surprised how good Argon got, didnt expect it.

这条是引用着另一条写的：x.com/sehozaifa/status/2111797472351916229