Gemini 3.1 Pro Preview
von Google DeepMind · Veroeffentlicht 2026-02-19
60.6
Durchschn. Score
$2.00/1M
Eingabepreis
$12.00/1M
Ausgabepreis
1.0M tokens (~524 books)
Kontextfenster
multimodal
Typ
Tested on 23 benchmarks with 60.6% average. Top scores: Chatbot Arena Elo — Overall (1492.6%), Chatbot Arena Elo — Coding (1455.7%), ARC-AGI (98.0%).
Benchmark-Ergebnisse
| Benchmark | Kategorie | Score | Bar |
|---|---|---|---|
| Chatbot Arena Elo — Overall | arena | 1492.6 | |
| Chatbot Arena Elo — Coding | arena | 1455.7 | |
| ARC-AGI | reasoning | 98.0 | |
| OTIS Mock AIME 2024-2025 | math | 95.6 | |
| GPQA diamond | knowledge | 92.1 | |
| Terminal Bench | coding | 78.4 | |
| SimpleQA Verified | knowledge | 77.3 | |
| ARC-AGI-2 | reasoning | 77.1 | |
| SWE-Bench verified | coding | 75.6 | |
| SimpleBench | reasoning | 75.5 | |
| WeirdML | coding | 72.1 | |
| MultiChallenge | knowledge | 71.4 | |
| MultiNRC | knowledge | 64.7 | |
| Artificial Analysis — Agentic Index | speed | 59.1 | |
| Artificial Analysis — Quality Index | speed | 57.2 | |
| Artificial Analysis — Coding Index | speed | 55.5 | |
| Chess Puzzles | knowledge | 55.0 | |
| FrontierMath-2025-02-28-Private | math | 36.9 | |
| APEX-Agents | agentic | 33.5 | |
| VisualToolBench (VTB) | knowledge | 29.0 | |
| PostTrainBench | knowledge | 21.6 | |
| EnigmaEval | knowledge | 19.8 | |
| FrontierMath-Tier-4-2025-07-01-Private | math | 16.7 |
Aehnliche Modelle
Alibaba
60.6
Alibaba Qwen
60.7
Google DeepMind
60.5
OpenAI
60.4