GPT-5.1-Codex-Max vs Kimi K2 Thinking
Lado a lado. Cada métrica. Cada benchmark.
| Tipo | GPT-5.1-Codex-Max | Kimi K2 Thinking |
|---|---|---|
| Provider |  OpenAI |  moonshotai |
| pontuação média | 72.0 | 53.3 |
| Preço de entrada | $1.25 | $0.60 |
| Preço de saída | $10.00 | $2.50 |
| Janela de contexto | 400K tokens (~200 books) | 262K tokens (~131 books) |
| Lançado em | 2025-12-04 | 2025-11-06 |
| Código aberto | Proprietary | Open Source |
Pontuações de benchmark
8 benchmarks · GPT-5.1-Codex-Max: 8, Kimi K2 Thinking: 0
| Benchmark | Categoria | GPT-5.1-Codex-Max | Kimi K2 Thinking |
|---|---|---|---|
| LiveBench — Agentic Coding | coding | 56.7 | 38.3 |
| LiveBench — Coding | coding | 81.4 | 67.4 |
| LiveBench — Data Analysis | reasoning | 54.9 | 52.3 |
| LiveBench — If | language | 67.1 | 62.0 |
| LiveBench — Language | language | 75.4 | 66.5 |
| LiveBench — Mathematics | math | 83.7 | 81.1 |
| LiveBench — Overall | knowledge | 72.0 | 61.6 |
| LiveBench — Reasoning | reasoning | 84.6 | 63.5 |