WHAT NEW IN X10.30 FRI

X 上的 AI 合集档案 · 22 个合集 · 60,637 条

gemini-4VIDEO

Gemini 4 Argon 多项基准超 Astra 与 Opus

视频列出 Argon 在法务任务 19.6%、DeepSWE、Vibe Code Bench、GraphWalks 领先,网络安全与 Astra 打平,定价 $2/$10。

Johan
@liebert_202628 天前收录原推发布于 2026.10.01
Johan @liebert_2026原推发布于 2026.10.01
在 X 上打开原推

HOLY FUCK GOOGLE IS BACK IN THE GAME gemini 4 argon: > beats astra and opus on work automation > 19.6% on legal tasks vs astra’s 5.4% and opus’s 3.8% > leads DeepSWE and Vibe Code Bench > beats both on long-context GraphWalks > ties astra at 68% on cybersecurity > $2 in / $10 out introductory pricing i owe sundar an apology

这条是引用着另一条写的:x.com/googledeepmind/status/2105388084154056939

x.com/liebert_2026/status/2105394187860619453采集于 2026.10.01 10:00