Atlas

Models

← All models

Gemini 2.5 Pro Preview 05-06

Google DeepMind · 2025-05-06

Gemini 2.5 Pro Preview 05-06 was Google's May 6, 2025 I/O edition update to Gemini 2.5 Pro, with stronger coding and function-calling capabilities. It accepts text, images, audio, and video, produces text, and supports up to 1,048,576 input tokens.

Benchmark scores

BenchmarkScore
Aider Polyglot76.9
Artificial Analysis Intelligence Index22.3
BRUMO 202589.2
BuseyBench SVG4.1
Capability144.6
Confabulations Leaderboard Confabulation Rate5.9
Confabulations Leaderboard Non-Response Rate15.3
Confabulations Leaderboard Weighted Score10.6
DeepResearchBench0.3
EnigmaEval2.4
Epoch Capabilities Index (ECI)143
EQ-Bench 31209
EQ-Bench 3 (Rubric Score)64.8
EQ-Bench 3: Analytic8.2
EQ-Bench 3: Assertive4.0
EQ-Bench 3: Compliant7.2
EQ-Bench 3: Empathy7.4
EQ-Bench 3: Humanlike5.8
EQ-Bench 3: Insight7.1
EQ-Bench 3: Moralising2.9
EQ-Bench 3: Pragmatic7.2
EQ-Bench 3: Safety5.9
EQ-Bench 3: Social IQ6.5
EQ-Bench 3: Warm7.7
Fiction.liveBench66.7
GeoBench ACW Country Accuracy86.0
Global MMLU Lite88.6
HMMT Feb 202580.8
Humanity's Last Exam17.8
Humanity's Last Exam (Text Only)18.4
Loading Atlas data…