Gemini 3 Pro (Low): benchmark results
Provider: Google. Released 2025-11-18. Access: API.
Unified ELO 1765 ± 23, rank #201 of 2133 rated models, from 24 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| BaFCo - Coarse Layout Analysis (CoT) | 25.13 | Mean average precision at IoU 0.3 (0-100): class-wise averag | 100 |
| GISA - List EM | 27.08 | Exact match (%) on the 48 list-type queries (an ordered list | 96.7 |
| GISA - Set F1 | 63.82 | F1 (%) between the predicted and gold sets on the 50 set-typ | 93.3 |
| Korean CSAT 2026 (Easy Mode) - English | 100 | Points (out of 100) | 90.5 |
| VitaBench | 30 | Cross-Scenario Avg@4 (%) | 90 |
| MCPMark | 50.79 | Pass@1 (%) | 89.5 |
| BaFCo - Coarse Layout Analysis (Zero-shot) | 25.78 | Mean average precision at IoU 0.3 (0-100): class-wise averag | 88.9 |
| BaFCo - Layout Analysis (CoT) | 8.53 | Mean average precision at IoU 0.3 (0-100): class-wise averag | 88.9 |
| BaFCo - Layout Analysis (Zero-shot) | 9 | Mean average precision at IoU 0.3 (0-100): class-wise averag | 88.9 |
| Korean CSAT 2026 (Easy Mode) - Chemistry I | 48 | Points (out of 50) | 80.1 |
| GISA - Table Item F1 | 64.93 | Item-level F1 (%) between predicted and gold table cells on | 80 |
| Korean CSAT 2026 (Easy Mode) - Physics I | 42 | Points (out of 50) | 78.7 |
Interactive version: theaggregate.ai/model?slug=gemini-3-pro-low · How It Works · Data refreshed daily, snapshot 2026-10-11.