Google says Gemini 4 Argon beats GPT-6 Astra and Opus 5.5 on 13 of 19 benchmarks

Sam Badawi @Sam_BadawiOriginally posted 2026.10.29

Open original on X

JUST IN: Google $GOOGL says its Gemini 4 Argon AI model outperforms OpenAI's GPT-6 Astra and Anthropic's Claude Opus 5.5 across most of its published benchmarks. Google's results show Argon leading 13 of 19 tests covering areas like coding, reasoning, and scientific tasks. Google is also reportedly testing a newer model called Gemini 4 Carbon internally, although official benchmarks for Carbon haven't been released.

This post quotes another post: “Google announces Gemini 4 Carbon with cybersecurity and long-task focus”