BenchAtlas

Rankings / Google DeepMind

Gemini 2.5 Flash Preview 04.17

released 2025-04-17

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

15 results
external indices
ECI140.6 pointsindependentEpoch AI Benchmarking Hub
ECIthinking · 24K budget140.6 pointsindependentEpoch AI Benchmarking Hub
ECIthinking · 16K budget140.6 pointsindependentEpoch AI Benchmarking Hub
knowledge science
GPQA Diamond
implementation=vals-ai
57.6 %independentVals AI
multimodal
MMMU
implementation=vals-ai
thinking71.9 %independentVals AI
MMMU
implementation=vals-ai
68.0 %independentVals AI
professional
CorpFin54.2 %independentVals AI
LegalBench83.8 %independentVals AI
MedQAthinking91.0 %independentVals AI
MedQA86.7 %independentVals AI
TaxEval71.2 %independentVals AI
TaxEvalthinking70.5 %independentVals AI
reasoning math
MATH-500
implementation=vals-ai
thinking91.8 %independentVals AI
MATH-500
implementation=vals-ai
91.6 %independentVals AI
OTIS Mock AIME 2024–202573.1 %independentEpoch AI Benchmarking Hub
2025-04-22

Related variants

Gemini 2.5 Flash Preview 05.20Gemini 2.5 Flash Preview 09 2025