GPT-5 Mini vs o4 Mini
Lado a lado. Cada métrica. Cada benchmark.
| Tipo | GPT-5 Mini | o4 Mini |
|---|---|---|
| Provider |  OpenAI | OpenAI |
| puntuación promedio | 56.0 | 53.2 |
| Precio de entrada | $0.25 | $1.10 |
| Precio de salida | $2.00 | $4.40 |
| Ventana de contexto | 400K tokens (~200 books) | 200K tokens (~100 books) |
| Publicado el | 2025-08-07 | 2025-04-16 |
| Código abierto | Proprietary | Proprietary |
Puntuaciones de benchmark
18 benchmarks · GPT-5 Mini: 10, o4 Mini: 7
| Benchmark | Categoría | GPT-5 Mini | o4 Mini |
|---|---|---|---|
| ARC-AGI | reasoning | 54.3 | 58.7 |
| ARC-AGI-2 | reasoning | 4.4 | 6.1 |
| Fiction.LiveBench | knowledge | 69.4 | 77.8 |
| FrontierMath-2025-02-28-Private | math | 27.2 | 24.8 |
| FrontierMath-Tier-4-2025-07-01-Private | math | 6.3 | 6.3 |
| GPQA diamond | knowledge | 66.7 | 72.8 |
| HELM — GPQA | knowledge | 75.6 | 73.5 |
| HELM — IFEval | language | 92.7 | 92.9 |
| HELM — MMLU-Pro | knowledge | 83.5 | 82.0 |
| HELM — Omni-MATH | math | 72.2 | 72.0 |
| HELM — WildBench | reasoning | 85.5 | 85.4 |
| HLE | knowledge | 15.4 | 13.9 |
| MATH level 5 | math | 97.8 | 97.8 |
| OTIS Mock AIME 2024-2025 | math | 86.7 | 81.7 |
| SimpleQA Verified | knowledge | 21.0 | 23.9 |
| SWE-Bench Verified (Bash Only) | coding | 59.8 | 45.0 |
| VPCT | knowledge | 10.3 | 36.3 |
| WeirdML | coding | 52.7 | 52.6 |