WHAT NEW IN X11.06 FRI

AI collections from X · 23 collections · 69,408 posts

gemini-4POST

Gemini 4 Argon benchmarks questioned over Gemini CLI vs Antigravity

A user argues that benchmarking Gemini 4 Argon via Gemini CLI instead of Antigravity makes the results untrustworthy.

FanXing
@fanxing321Archived 17 days agoOriginally posted 2026.10.18
FanXing @fanxing321Originally posted 2026.10.18
Open original on X

我不理解为什么这些基准测试他们在测试 Gemini 4 Argon 时,总是要使用 Gemini CLI,而不是 antigravity。 这导致这个基准测试变得不可信

This post quotes another post: “Gemini 4 Argon (high) scores 20.2% on WeirdML”

x.com/fanxing321/status/2111800384381763849Archived 2026.10.19 10:00