Gemini 2.5 Pro: benchmark results
Google Gemini 2.5 Pro model for advanced reasoning and multimodal tasks. Provider: Google. Released 2025-03-25. Access: API.
Unified ELO 1661 ± 1, rank #109 of 1392 rated models, from 1369 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AI for Education Visual Maths - Algebra | 100 | Accuracy (%) | 100 |
| Active Evidence-Seeking and Diagnostic Reasoni | 49.4 | Task 2: Active Seeking Exact Accuracy (self-reported) | 100 |
| BIRD-Interact (c-Interact) | 20.92 | Normalized Reward | 100 |
| BesiegeField - Car | 34.34 | Best-scaffold Mean Score | 100 |
| BesiegeField - Catapult | 9.83 | Best-scaffold Mean Score | 100 |
| Chartographer | 94.7 | ChartQA OA (self-reported) | 100 |
| DeepResearch Bench - Effective Citations | 165.34 | Effective Citations | 100 |
| Episodic Memory | 96.8 | Simple Recall (%) | 100 |
| EuroEval Czech | 70.02 | Average Score (%) | 100 |
| EuroEval Estonian | 62.38 | Average Score (%) | 100 |
| EuroEval Hungarian | 67.51 | Average Score (%) | 100 |
| EuroEval Norwegian Knowledge | 87.61 | Knowledge Average Score (%) | 100 |
Interactive version: theaggregate.ai/model?slug=gemini-2-5-pro · How It Works · Data refreshed daily, snapshot 2026-09-05.