Rankings / Google DeepMind

Gemini 3 Deep Think 022026

released 2026-02-12

BenchAtlas Index

as of 2026-07-14

Not enough benchmark coverage for a composite score (needs ≥4 families across ≥3 categories). The per-benchmark evidence below stands on its own — sparse coverage is an evidence gap, not a low score.

Benchmark evidence

4 results

reasoning math

ARC-AGI-1older version split=semi_private · model_type=CoT	96.0 %	independent	ARC Prize Leaderboard
ARC-AGI-2 split=semi_private · model_type=CoT	84.6 %	independent	ARC Prize Leaderboard
ARC-AGI-2 split=semi_private · model_type=CoT	13.6 usd_per_task	independent	ARC Prize Leaderboard
ARC-AGI-1older version split=semi_private · model_type=CoT	7.2 usd_per_task	independent	ARC Prize Leaderboard