Gemini 3.8 Flash (Medium): benchmark results

Provider: Google. Released 2026-09-02. Access: API.

Unified ELO 1735 ± 1, rank #47 of 1761 rated models, from 22 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
AA Long Context Reasoning84Accuracy (%)98.7
AA GPQA Diamond93.54Accuracy (%)97.3
AA Omniscience - Law59.5Accuracy (%)96.7
AA Omniscience28.58Score96.1
Artificial Analysis Intelligence Index46.77Intelligence Index95.3
AA MMMU-Pro84.22Accuracy (%)95
Tau3 Banking45.77Success Rate (%)94.3
AA Omniscience - Business44Accuracy (%)93.7
AA Omniscience - Humanities & Social Sciences52.6Accuracy (%)93.6
AA-Omniscience Accuracy52.98Accuracy (%)93.6
AA Humanity's Last Exam42.12Accuracy (%)92.8
Computer Anthology Terminal Tasks (Terminus-2)51.8pass@1 (%)92.3

Interactive version: theaggregate.ai/model?slug=gemini-3-8-flash-medium · How It Works · Data refreshed daily, snapshot 2026-09-05.