AGIdex
agidex/Technologies/Gemini 3.1 Pro
AI MODEL

Gemini 3.1 Pro

CLOSED
RELEASED 2026-02

Google DeepMind's current flagship. 1M token context. 77.1% on ARC-AGI-2, 94.3% GPQA Diamond.

Benchmark coverage · per sub-capability

Each score is anchored to human baseline (100). Source URLs link to the original benchmark, leaderboard, or release note.

№ 04ReasoningCOGNITIVE
Sub-capabilityQualitySourceScore vs. baselineScore
Abstract Reasoningindependentarcprize.org/leaderboard
77%
Logical Reasoningvendor reporteddeepmind.google/models/model-cards/gemini-3-1-pro
71%
№ 01CommunicationCOGNITIVE
Sub-capabilityQualitySourceScore vs. baselineScore
Multilingualvendor reporteddeepmind.google/models/model-cards/gemini-3-1-pro
110%
№ 02VisionCOGNITIVE
Sub-capabilityQualitySourceScore vs. baselineScore
Visual Reasoningvendor reporteddeepmind.google/models/model-cards/gemini-3-1-pro
80%