Gemini 3.1 Pro (Preview) (Medium): benchmark results
Provider: Google. Released 2026-02-19. Access: API.
Aggregate Rating 1767 ± 11, rank #204 of 2170 rated models, from 107 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| BabelArena - Tau3-Bench | 71.4 | Pass^1 (%; the text component of tau3-Bench (simulated-user | 100 |
| Blind-Spots-Bench - Multimodal-to-Text | 66.9 | Accuracy (%; mean@4 over four samples on the questions with | 100 |
| Blind-Spots-Bench - Text-only (Code Execution) | 86.11 | Accuracy (%; mean@4 on the text-only questions with a Python | 100 |
| CP-SynC-XL - MiniZinc + OR-Tools | 54.4 | Correct instances (%; share of the 4,577 CP-SynC-XL instance | 100 |
| CP-SynC-XL - Python (Heuristic Prompt) | 64 | Correct instances (%; share of the 4,577 CP-SynC-XL instance | 100 |
| CP-SynC-XL - Python + OR-Tools | 71.6 | Correct instances (%; share of the 4,577 CP-SynC-XL instance | 100 |
| CP-SynC-XL - Python + OR-Tools (Heuristic Prompt) | 72.2 | Correct instances (%; share of the 4,577 CP-SynC-XL instance | 100 |
| MMGist - Diagram and OCR | 82.3 | Accuracy (%) on the Diagram and OCR dimension (items kept fr | 100 |
| MMGist - Expert Knowledge | 70.2 | Accuracy (%) on the Expert Knowledge dimension (items kept f | 100 |
| MMGist - Medical | 58 | Accuracy (%) on the Medical dimension (items kept from SLAKE | 100 |
| MMGist - Spatial Understanding | 79.7 | Accuracy (%) on the Spatial Understanding dimension (items k | 100 |
| MMGist - Visual Logic | 40.4 | Accuracy (%) on the Visual Logic dimension (items kept from | 100 |
Interactive version: theaggregate.ai/model?slug=gemini-3-1-pro-preview-medium · How It Works · Data refreshed daily, snapshot 2026-11-05.