Gemini 2.5 Flash (Preview 05-20) — benchmark results
May 20 Gemini 2.5 Flash preview snapshot. Provider: Google. Released 2025-06-17. Access: API.
Unified ELO 1728 ± 30, rank #223 of 1776 rated models, from 36 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| DuckDB-NSQL | 77.3 | Execution Accuracy (%) | 98.4 |
| KOFFVQA - Document Understanding | 100 | Score (%) | 98.1 |
| KOFFVQA - Table Understanding | 96.67 | Score (%) | 98.1 |
| KOFFVQA - Korean Recognition | 90 | Score (%) | 95.7 |
| MMMU Benchmark | 79.7 | Validation Score | 95.2 |
| KOFFVQA - Commonsense Reasoning | 92 | Score (%) | 95.1 |
| KOFFVQA | 86.76 | Overall Score (%) | 92.6 |
| KOFFVQA - Graph and Chart Understanding | 88 | Score (%) | 92 |
| Context Arena MRCR (2-needle) | 78.3 | AUC@1M (%) | 90 |
| KOFFVQA - Recognition | 90 | Score (%) | 87 |
| RepairBench | 40.6 | Plausible@1 (%) | 84.2 |
| KOFFVQA - Relationship | 73.33 | Score (%) | 84 |
Interactive version: theaggregate.ai/model?slug=gemini-2-5-flash-preview-05-20 · How the rankings work · Data refreshed daily, snapshot 2026-07-22.