Gemini 2.5 Flash vs GPT-5 Mini
Lado a lado. Cada métrica. Cada benchmark.
| Tipo | Gemini 2.5 Flash | GPT-5 Mini |
|---|---|---|
| Provider |  Google DeepMind |  OpenAI |
| pontuação média | 40.0 | 56.0 |
| Preço de entrada | $0.30 | $0.25 |
| Preço de saída | $2.50 | $2.00 |
| Janela de contexto | 1.0M tokens (~524 books) | 400K tokens (~200 books) |
| Lançado em | 2025-06-17 | 2025-08-07 |
| Código aberto | Proprietary | Proprietary |
Pontuações de benchmark
15 benchmarks · Gemini 2.5 Flash: 0, GPT-5 Mini: 15
| Benchmark | Categoria | Gemini 2.5 Flash | GPT-5 Mini |
|---|---|---|---|
| ARC-AGI | reasoning | 32.3 | 54.3 |
| ARC-AGI-2 | reasoning | 2.5 | 4.4 |
| Fiction.LiveBench | knowledge | 47.2 | 69.4 |
| FrontierMath-2025-02-28-Private | math | 4.8 | 27.2 |
| FrontierMath-Tier-4-2025-07-01-Private | math | 4.2 | 6.3 |
| HELM — GPQA | knowledge | 39.0 | 75.6 |
| HELM — IFEval | language | 89.8 | 92.7 |
| HELM — MMLU-Pro | knowledge | 63.9 | 83.5 |
| HELM — Omni-MATH | math | 38.4 | 72.2 |
| HELM — WildBench | reasoning | 81.7 | 85.5 |
| HLE | knowledge | 7.7 | 15.4 |
| OTIS Mock AIME 2024-2025 | math | 73.0 | 86.7 |
| Terminal Bench | coding | 17.1 | 34.8 |
| VPCT | knowledge | 7.0 | 10.3 |
| WeirdML | coding | 41.0 | 52.7 |