GPT-4.1 Mini vs GPT-5 Mini
Côte à côte. Chaque métrique. Chaque benchmark.
| Type | GPT-4.1 Mini | GPT-5 Mini |
|---|---|---|
| Provider |  OpenAI | OpenAI |
| score moyen | 44.5 | 56.0 |
| Prix d'entrée | $0.40 | $0.25 |
| Prix de sortie | $1.60 | $2.00 |
| Fenêtre de contexte | 1.0M tokens (~524 books) | 400K tokens (~200 books) |
| Sorti le | 2025-04-14 | 2025-08-07 |
| Code source ouvert | Proprietary | Proprietary |
Scores de benchmark
14 benchmarks · GPT-4.1 Mini: 0, GPT-5 Mini: 14
| Benchmark | Catégorie | GPT-4.1 Mini | GPT-5 Mini |
|---|---|---|---|
| ARC-AGI | reasoning | 3.5 | 54.3 |
| ARC-AGI-2 | reasoning | 0.1 | 4.4 |
| Fiction.LiveBench | knowledge | 44.4 | 69.4 |
| FrontierMath-2025-02-28-Private | math | 4.5 | 27.2 |
| GPQA diamond | knowledge | 54.5 | 66.7 |
| HELM — GPQA | knowledge | 61.4 | 75.6 |
| HELM — IFEval | language | 90.4 | 92.7 |
| HELM — MMLU-Pro | knowledge | 78.3 | 83.5 |
| HELM — Omni-MATH | math | 49.1 | 72.2 |
| HELM — WildBench | reasoning | 83.8 | 85.5 |
| MATH level 5 | math | 87.3 | 97.8 |
| OTIS Mock AIME 2024-2025 | math | 44.7 | 86.7 |
| SWE-Bench Verified (Bash Only) | coding | 23.9 | 59.8 |
| WeirdML | coding | 37.6 | 52.7 |