GPT-4.1 Nano
来自 OpenAI · 发布于 2025-04-14
35.2
平均分
$0.10/1M
输入价格
$0.40/1M
输出价格
1.0M tokens (~524 books)
上下文窗口
multimodal
类型
Tested on 14 benchmarks with 35.2% average. Top scores: HELM — IFEval (84.3%), HELM — WildBench (81.1%), MATH level 5 (70.0%).
基准测试分数
| 基准测试 | 类别 | 分数 | Bar |
|---|---|---|---|
| HELM — IFEval | language | 84.3 | |
| HELM — WildBench | reasoning | 81.1 | |
| MATH level 5 | math | 70.0 | |
| HELM — MMLU-Pro | knowledge | 55.0 | |
| HELM — GPQA | knowledge | 50.7 | |
| HELM — Omni-MATH | math | 36.7 | |
| GPQA diamond | knowledge | 31.9 | |
| OTIS Mock AIME 2024-2025 | math | 28.8 | |
| Fiction.LiveBench | knowledge | 25.0 | |
| WeirdML | coding | 19.0 | |
| Aider polyglot | coding | 8.9 | |
| FrontierMath-2025-02-28-Private | math | 1.0 | |
| ARC-AGI-2 | reasoning | 0.1 | |
| ARC-AGI | reasoning | 0.1 |