Gemini 3.5 Flash: benchmark results
Google's speed-oriented Gemini 3.5 Flash model, balancing capability with lower latency. Provider: Google. Released 2026-05-19. Access: API.
Aggregate Rating 1809 ± 3, rank #60 of 1656 rated models, from 1328 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AA APEX-Agents | 47.05 | Pass@1 (%) | 100 |
| APEX-Agents | 66.1 | Mean Score (ReAct) (imported) | 100 |
| AeroCopilotBench - Tier-2 Ineffective-Action Rate | 5.5 | Ineffective write attempts (%; lower is better) | 100 |
| AeroCopilotBench - Tier-2 Safety Compliance Rate | 99.5 | Episodes without a hard safety violation (%) | 100 |
| Arabic Broad Leaderboard | 9.25 | Average Score (0-10) | 100 |
| ArtECulture | 49.88 | Accuracy (%; mean of the English, Chinese and Arabic culture | 100 |
| CHARM - Boundary Compliance | 64.4 | Accuracy (%) | 100 |
| Chess APO Bench | 55.4 | Puzzle accuracy (%) on the 559-puzzle Lichess test split wit | 100 |
| Computer Anthology Terminal Tasks (Gemini CLI) | 24.4 | pass@1 (%) | 100 |
| DRNoise | 26 | Noisy-condition accuracy (%; 100 deep-research tasks whose u | 100 |
| DRNoise (Clean) | 96.3 | Clean-condition accuracy (%; 100 deep-research tasks whose u | 100 |
| ELBench - Basic Education | 94.6 | Module score (%; mean rubric score on 45 ELMES teaching scen | 100 |
Interactive version: theaggregate.ai/model?slug=gemini-3-5-flash · How It Works · Data refreshed daily, snapshot 2026-10-20.