GPT-5.1-Codex-Max
来自 OpenAI · 发布于 2025-12-04
72.0
平均分
$1.25/1M
输入价格
$10.00/1M
输出价格
400K tokens (~200 books)
上下文窗口
multimodal
类型
Tested on 8 benchmarks with 72.0% average. Top scores: LiveBench — Reasoning (84.6%), LiveBench — Mathematics (83.7%), LiveBench — Coding (81.4%).
基准测试分数
| 基准测试 | 类别 | 分数 | Bar |
|---|---|---|---|
| LiveBench — Reasoning | reasoning | 84.6 | |
| LiveBench — Mathematics | math | 83.7 | |
| LiveBench — Coding | coding | 81.4 | |
| LiveBench — Language | language | 75.4 | |
| LiveBench — Overall | knowledge | 72.0 | |
| LiveBench — If | language | 67.1 | |
| LiveBench — Agentic Coding | coding | 56.7 | |
| LiveBench — Data Analysis | reasoning | 54.9 |