Gemini 2.5 Deep Think: benchmark results
Provider: Google. Access: API.
Aggregate Rating 1769 ± 34, rank #146 of 10916 rated models, from 4 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| IUMB | 70.8 | Score (%) | 81.5 |
| FrontierMath - Tiers 1-3 | 29 | Accuracy (%, 290 problems) | 75.8 |
| FrontierMath - Tier 4 | 10.4 | Accuracy (%, 48 problems) | 64.8 |
| VoxelBench | 1468 | Rating | 30.4 |
Interactive version: theaggregate.ai/model?slug=gemini-2-5-deep-think · How It Works · Data refreshed daily, snapshot 2026-11-08.