GPT-5.1 vs GPT-5 Nano
Côte à côte. Chaque métrique. Chaque benchmark.
| Type | GPT-5.1 | GPT-5 Nano |
|---|---|---|
| Provider |  OpenAI | OpenAI |
| score moyen | 49.6 | 45.3 |
| Prix d'entrée | $1.25 | $0.05 |
| Prix de sortie | $10.00 | $0.40 |
| Fenêtre de contexte | 400K tokens (~200 books) | 400K tokens (~200 books) |
| Sorti le | 2025-11-13 | 2025-08-07 |
| Code source ouvert | Proprietary | Proprietary |
Scores de benchmark
16 benchmarks · GPT-5.1: 13, GPT-5 Nano: 3
| Benchmark | Catégorie | GPT-5.1 | GPT-5 Nano |
|---|---|---|---|
| ARC-AGI | reasoning | 72.8 | 20.7 |
| ARC-AGI-2 | reasoning | 17.6 | 2.6 |
| FrontierMath-2025-02-28-Private | math | 31.0 | 8.3 |
| FrontierMath-Tier-4-2025-07-01-Private | math | 12.5 | 2.1 |
| GPQA diamond | knowledge | 83.5 | 59.3 |
| HELM — GPQA | knowledge | 44.2 | 67.9 |
| HELM — IFEval | language | 93.5 | 93.2 |
| HELM — MMLU-Pro | knowledge | 57.9 | 77.8 |
| HELM — Omni-MATH | math | 46.4 | 54.7 |
| HELM — WildBench | reasoning | 86.3 | 80.6 |
| OTIS Mock AIME 2024-2025 | math | 88.6 | 81.1 |
| SimpleQA Verified | knowledge | 48.9 | 12.2 |
| SWE-Bench Verified (Bash Only) | coding | 66.0 | 34.8 |
| Terminal Bench | coding | 47.6 | 11.5 |
| VPCT | knowledge | 38.0 | 5.8 |
| WeirdML | coding | 60.8 | 38.1 |