Gemini 3.1 Pro (Preview) (High) — benchmark results
Gemini 3.1 Pro (Preview) evaluated at the high reasoning-effort setting. Provider: Google. Released 2026-02-19. Access: API.
Unified ELO 1986 ± 37, rank #27 of 1776 rated models, from 47 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| FutureEval | 19.84 | Unified Forecasting Score | 100 |
| MedCode | 59.06 | Score (self-reported) | 100 |
| SEAL - Humanity's Last Exam | 46.44 | Score | 100 |
| UGI - Natural Intelligence | 74.94 | NatInt Score | 99.3 |
| Chess Bench LLM | 2066 | Lichess Rating | 99 |
| NonoBench | 63.3 | Overall Accuracy (%) | 98.9 |
| LLM Chess (Saplin) | 1511.4 | ELO | 98.6 |
| UGI - Writing | 68.22 | Writing Score | 98.1 |
| Context Arena MRCR (2-needle) | 87.8 | AUC@1M (%) | 96.2 |
| LisanBench | 0.45 | Mean Path Length / Current Maximum | 96 |
| ATLAS | 68.52 | ATLAScore (self-reported) | 95.5 |
| Context Arena MRCR (8-needle) | 47.5 | AUC@1M (%) | 95.5 |
Interactive version: theaggregate.ai/model?slug=gemini-3-1-pro-preview-high · How the rankings work · Data refreshed daily, snapshot 2026-07-22.