Gemini 3 Pro vs Gemini 3 Flash Preview
Lado a lado. Cada métrica. Cada benchmark.
| Tipo | Gemini 3 Pro | Gemini 3 Flash Preview |
|---|---|---|
| Provider |  Google DeepMind | Google DeepMind |
| pontuação média | 60.5 | 49.1 |
| Preço de entrada | - | $0.50 |
| Preço de saída | - | $3.00 |
| Janela de contexto | - | 1.0M tokens (~524 books) |
| Lançado em | 2024-01-01 | 2025-12-17 |
| Código aberto | Proprietary | Proprietary |
Pontuações de benchmark
21 benchmarks · Gemini 3 Pro: 12, Gemini 3 Flash Preview: 9
| Benchmark | Categoria | Gemini 3 Pro | Gemini 3 Flash Preview |
|---|---|---|---|
| APEX-Agents | agentic | 18.4 | 24.0 |
| ARC-AGI | reasoning | 75.0 | 21.5 |
| ARC-AGI-2 | reasoning | 31.1 | 33.6 |
| Artificial Analysis — Agentic Index | speed | 45.0 | 49.7 |
| Artificial Analysis — Coding Index | speed | 39.4 | 42.6 |
| Artificial Analysis — Quality Index | speed | 41.3 | 46.4 |
| Chatbot Arena Elo — Coding | arena | 1437.6 | 1436.4 |
| Chatbot Arena Elo — Overall | arena | 1486.2 | 1473.9 |
| Chess Puzzles | knowledge | 31.0 | 38.0 |
| FrontierMath-2025-02-28-Private | math | 37.6 | 35.6 |
| FrontierMath-Tier-4-2025-07-01-Private | math | 18.8 | 4.2 |
| GeoBench | knowledge | 84.0 | 88.0 |
| GPQA diamond | knowledge | 90.2 | 77.6 |
| GSO-Bench | coding | 18.6 | 9.8 |
| OTIS Mock AIME 2024-2025 | math | 91.4 | 92.8 |
| SimpleBench | reasoning | 71.7 | 53.3 |
| SimpleQA Verified | knowledge | 72.9 | 67.4 |
| SWE-Bench verified | coding | 72.9 | 75.4 |
| Terminal Bench | coding | 69.4 | 64.3 |
| VPCT | knowledge | 86.5 | 58.9 |
| WeirdML | coding | 69.9 | 61.6 |