Gemini 3 Flash (High) — benchmark results
Gemini 3 Flash evaluated at the high reasoning-effort setting. Provider: Google. Released 2025-12-01. Access: API.
Unified ELO 1823 ± 27, rank #121 of 1776 rated models, from 34 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| VitaBench | 32.5 | Cross-Scenario Avg@4 (%) | 100 |
| LLM2014 Logic 2026-01 | 73.32 | Median Score | 95.5 |
| LLM2014 Logic 2025-12 | 76.9 | Median Score | 94 |
| LLM2014 Logic 2026-03 | 65.29 | Median Score | 90.2 |
| LLM2014 Logic 2026-02 | 66.09 | Median Score | 86.7 |
| Pencil Puzzle Bench - Mashu | 6.7 | Direct-ask Success Rate (%) | 82 |
| Pencil Puzzle Bench - Nurimisaki | 13.3 | Direct-ask Success Rate (%) | 81 |
| Pencil Puzzle Bench - Tapa | 13.3 | Direct-ask Success Rate (%) | 81 |
| LLM2014 Logic 2026-04 | 58.87 | Median Score | 80 |
| Pencil Puzzle Bench - LITS | 6.7 | Direct-ask Success Rate (%) | 78 |
| Pencil Puzzle Bench - Light Up | 6.7 | Direct-ask Success Rate (%) | 71 |
| APEX v1 Investment Banking | 59.8 | Score (%) | 69.4 |
Interactive version: theaggregate.ai/model?slug=gemini-3-flash-high · How the rankings work · Data refreshed daily, snapshot 2026-07-22.