ARC-AGI-3

Latest news

1. 16 Oct
   Google DeepMind’s Gemini 4 Argon scores 97.5% on ARC-AGI-3 benchmark

   Google DeepMind’s new Gemini 4 Argon model scored 97.5% on the ARC-AGI-3 benchmark using a vendor-specific test setup called the Provider Adapter Harness. The result makes DeepMind the second lab after OpenAI to pass 95%, though the score depends heavily on the testing method and Argon lags rivals in math and software engineering according to Epoch AI.

   Models · 1 source

   

Articles

No articles about ARC-AGI-3 yet.