<p>BREAKING 🚨: GEMINI 3.1 PRO HITS 77.1% ON ARC-AGI-2 AND 44.4%. ON HLE. ON PAR WITH OPUS 4.6 ON SWE BENCH VERIFIED.

ABSOLUTE SOTA 👀 https://t.co/g4BQdi2cui https://t.co/r9KVTiNcbM</p>

unknown

Gemini 3.1 Pro is here.

We’ve significantly improved the model’s overall intelligence so it can solve tougher problems. 🧵 https://t.co/upyJv1p7MA

TestingCatalog News 🗞

Gemini 3.1 Pro reportedly achieves 77.1% on ARC-AGI-2 and 44.4% on HLE benchmarks, claiming state-of-the-art performance and matching Claude Opus 4.6 on SWE-Bench Verified.

BREAKING 🚨: GEMINI 3.1 PRO HITS 77.1% ON ARC-AGI-2 AND 44.4%. ON HLE. ON PAR WITH OPUS 4.6 ON SWE BENCH VERIFIED.

ABSOLUTE SOTA 👀 https://t.co/g4BQdi2cui https://t.co/r9KVTiNcbM