Gemini 1.5 Pro (002): benchmark results
Gemini 1.5 Pro 002 stable API snapshot. Provider: Google. Released 2024-09-24. Access: API.
Unified ELO 1571 ± 1, rank #356 of 1392 rated models, from 964 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| CapArena-Auto | 56.2 | CapArena-Auto Score (avg) | 100 |
| CapArena-Auto vs CogVLM-19B | 61 | CapArena-Auto Score (vs CogVLM-19B) | 100 |
| CapArena-Auto vs GPT-4o | 29 | CapArena-Auto Score (vs GPT-4o) | 100 |
| CapArena-Auto vs MiniCPM-8B | 78.5 | CapArena-Auto Score (vs MiniCPM-8B) | 100 |
| HELM Finance - FinQA | 58.7 | Program Accuracy | 100 |
| MEGA-Bench Task - Actor Recognition In Movie | 100 | Task Score (%) | 100 |
| MEGA-Bench Task - Autorater Mask | 85.7 | Task Score (%) | 100 |
| MEGA-Bench Task - Constrained Generation Contain Position Length | 93.3 | Task Score (%) | 100 |
| MEGA-Bench Task - Constrained Generation Multi Contain Repeat | 6.7 | Task Score (%) | 100 |
| MEGA-Bench Task - Counting Multi Image | 71.4 | Task Score (%) | 100 |
| MEGA-Bench Task - FigureQA | 71.4 | Task Score (%) | 100 |
| MEGA-Bench Task - GUI Act Web Multi | 47 | Task Score (%) | 100 |
Interactive version: theaggregate.ai/model?slug=gemini-1-5-pro-002 · How It Works · Data refreshed daily, snapshot 2026-09-05.