Report
Gemini 4 Argon is claimed to be in a close tie with GPT-6-Astra on RSI-Inference Bench
A post calls it the first model outside Anthropic and OpenAI to pass a Day-0 Model Support task.
TLDR
In an October 26 post, someone describing RSI-Inference Bench as “our” benchmark said Google DeepMind’s Gemini 4 Argon, with Antigravity, had made a significant leap on the bench and was in a close tie with GPT-6-Astra. The post also claimed it was the first model outside Anthropic and OpenAI to pass a Day-0 Model Support task.
Combined views
44.1K
3 Sources, first seen ago

