Gemini 3.1 Pro Preview vs GPT-5.2
Lado a lado. Cada métrica. Cada benchmark.
| Tipo | Gemini 3.1 Pro Preview | GPT-5.2 |
|---|---|---|
| Provider |  Google DeepMind |  OpenAI |
| puntuación promedio | 60.6 | 54.0 |
| Precio de entrada | $2.00 | $1.75 |
| Precio de salida | $12.00 | $14.00 |
| Ventana de contexto | 1.0M tokens (~524 books) | 400K tokens (~200 books) |
| Publicado el | 2026-02-19 | 2025-12-10 |
| Código abierto | Proprietary | Proprietary |
Puntuaciones de benchmark
16 benchmarks · Gemini 3.1 Pro Preview: 11, GPT-5.2: 5
| Benchmark | Categoría | Gemini 3.1 Pro Preview | GPT-5.2 |
|---|---|---|---|
| APEX-Agents | agentic | 33.5 | 34.3 |
| ARC-AGI | reasoning | 98.0 | 86.2 |
| ARC-AGI-2 | reasoning | 77.1 | 52.9 |
| Chatbot Arena Elo — Coding | arena | 1455.7 | 1403.1 |
| Chatbot Arena Elo — Overall | arena | 1492.6 | 1439.5 |
| Chess Puzzles | knowledge | 55.0 | 49.0 |
| FrontierMath-2025-02-28-Private | math | 36.9 | 40.7 |
| FrontierMath-Tier-4-2025-07-01-Private | math | 16.7 | 18.8 |
| GPQA diamond | knowledge | 92.1 | 88.5 |
| OTIS Mock AIME 2024-2025 | math | 95.6 | 96.1 |
| PostTrainBench | knowledge | 21.6 | 21.4 |
| SimpleBench | reasoning | 75.5 | 35.0 |
| SimpleQA Verified | knowledge | 77.3 | 38.9 |
| SWE-Bench verified | coding | 75.6 | 73.8 |
| Terminal Bench | coding | 78.4 | 64.9 |
| WeirdML | coding | 72.1 | 72.2 |