WHAT NEW IN X11.01 SUN

X 上的 AI 合集档案 · 22 个合集 · 63,170 条

gemini-4IMAGE

用户称 Gemini 4 Argon 是年度最差发布之一,幻觉率超 80%

Theo 称 Gemini 在评审中 80% 以上出现幻觉,并把 Gemini 4 Argon 列为年度最差发布之一;帖中提及未发布的内部模型 Carbon。

vorpal
@vorpal_onchain4 天前收录原推发布于 2026.10.26
用户称 Gemini 4 Argon 是年度最差发布之一,幻觉率超 80%
vorpal @vorpal_onchain原推发布于 2026.10.26
在 X 上打开原推

Theo says Gemini hallucinates in over 80% of his review runs he added a "Gemini the jester" section to his council skill, where models review each other's work, and Gemini keeps making things up he's also called Gemini 4 Argon one of the worst releases of the year this is why I'm not buying the $200 Gemini plan Google's answer might be Carbon, an internal checkpoint reportedly matching Opus 5.5 in coding it isn't released yet can Carbon fix this, or is Google cooked?

x.com/vorpal_onchain/status/2114576473625432465采集于 2026.10.27 09:53