Gemini 2.0 Flash (Thinking): benchmark results
Gemini 2.0 Flash evaluated with thinking enabled. Provider: Google. Released 2024-12-11. Access: API.
Unified ELO 1573 ± 1, rank #544 of 1761 rated models, from 54 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Video-Holmes | 43.1 | Overall (self-reported) | 93.3 |
| SEAL - Arabic | 1120 | Score | 86.1 |
| SEAL - Spanish | 1108 | Score | 81 |
| EmoBench-M | 60.6 | Average Score | 80 |
| VRBench | 60.97 | Overall (self-reported) | 78.6 |
| SEAL - Coding | 1108 | Score | 77.8 |
| LongBench v2 | 56 | Accuracy (%) | 76.5 |
| SEAL - Agentic Tool Use (Chat) | 57.36 | Score | 75.8 |
| WebApp1K | 85.9 | Pass@1 (%) | 72.7 |
| Divergent Thinking | 4.41 | Divergence Score | 69.4 |
| MMToM-QA | 54 | All (%) | 69.2 |
| Episodic Memory | 70.8 | Simple Recall (%) | 67.5 |
Interactive version: theaggregate.ai/model?slug=gemini-2-0-flash-thinking · How It Works · Data refreshed daily, snapshot 2026-09-05.