Gemini 3.1 Flash (Preview): benchmark results

Provider: Google. Released 2026-03-03. Access: API.

Aggregate Rating 1658 ± 42, rank #569 of 10916 rated models, from 6 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
MapSatisfyBench - Factual Faithfulness94.48Factual-faithfulness rate (%): share of required factual cla85.7
AudioMC36.06Score (imported)70.4
MapSatisfyBench - Acceptance Rate52.31Accepted-response score (%): per instance the explicit-compl21.4
MapSatisfyBench - Explicit Completion85.73Explicit-factor completion rate (%): share of the query's st21.4
MapSatisfyBench - Implicit Intent Satisfaction58.34Implicit-intent satisfaction rate (%): evidence-weighted sat21.4
MapSatisfyBench - Tool Selection42.06Tool-selection score (%): set overlap between the valid tool21.4

Interactive version: theaggregate.ai/model?slug=gemini-3-1-flash-preview · How It Works · Data refreshed daily, snapshot 2026-11-08.