Gemini 1.5 Flash (002): benchmark results
Gemini 1.5 Flash 002 stable API snapshot. Provider: Google. Released 2024-09-24. Access: API.
Unified ELO 1518 ± 1, rank #609 of 1392 rated models, from 865 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MEGA-Bench Task - NLVR2 Two Image Compare QA | 85.7 | Task Score (%) | 98.8 |
| MEGA-Bench Task - Video Content Reasoning | 88.9 | Task Score (%) | 98.8 |
| MEGA-Bench Task - Webpage Code Understanding | 88.9 | Task Score (%) | 98.8 |
| OpenVLM MMBench V1.1 CN - Identity Reasoning | 100 | Accuracy (%) | 97.9 |
| MEGA-Bench Task - Electricity Plot Future Prediction | 90.2 | Task Score (%) | 97.7 |
| MEGA-Bench Task - LLaVAGuard | 78.6 | Task Score (%) | 97.7 |
| MEGA-Bench Task - Medical Retrieval Given Surgeon Activity | 57.1 | Task Score (%) | 97.7 |
| MEGA-Bench Task - Video Camera Motion Description | 35.7 | Task Score (%) | 97.7 |
| MEGA-Bench Task - MMSoc Misinformation GossipCop | 78.6 | Task Score (%) | 96.5 |
| MEGA-Bench Task - Multi Load Type Prediction From Plot | 53.6 | Task Score (%) | 96.5 |
| OpenVLM MMMU - History | 83.3 | Accuracy (%) | 96.5 |
| OpenVLM MMBench V1.1 EN - Function Reasoning | 96.7 | Accuracy (%) | 96.3 |
Interactive version: theaggregate.ai/model?slug=gemini-1-5-flash-002 · How It Works · Data refreshed daily, snapshot 2026-09-05.