Gemini 2.5 Pro (Preview) — benchmark results
Google preview Gemini 2.5 Pro model, an early Pro snapshot from the Gemini 2.5 family. Provider: Google. Released 2025-03-25. Access: API.
Unified ELO 1715 ± 42, rank #301 of 1806 rated models, from 8 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MATH-Perturb (Hard) | 88.89 | Accuracy (%) | 100 |
| PlatinumBench (MIT) | 0.53 | Avg Error Rate (%) | 100 |
| Gapminder AI Worldview | 93.1 | Correct Rate (%) | 97.1 |
| EQ-Bench Creative Writing v3 | 1281.8 | Elo | 60.6 |
| SciArena | 1009.7 | Elo Rating | 51.4 |
| GosuEvals | 58.3 | Score (%) | 44.4 |
| Reward Hacking Benchmark (RHB) | 4.6 | Exploit rate (%) (self-reported) | 41.7 |
| SEAL - SWE-Bench Pro (Private Dataset) | 10.14 | Score | 15.4 |
Interactive version: theaggregate.ai/model?slug=gemini-2-5-pro-preview · How It Works · Data refreshed daily, snapshot 2026-08-07.