跳轉到主要內容

GPT-4.1 nano

OpenAIGPTProprietary

描述

GPT-4.1 nano is OpenAI's fastest and cheapest model available in the GPT-4.1 family. It delivers exceptional performance at a small size with its 1 million token context window. Ideal for tasks like classification or autocompletion.

發布日期
2025-04-14
參數規模
上下文長度
1.0M
支援模態
file, image, text

能力雷達圖

27
general
20
coding
36
reasoning
33
science估算
10
agents
85
multimodal

Science 在缺少專門科學評測時使用推理能力代理估算。

排行榜排名

領域#排名分數來源
代码能力榜341
20.0
AA
通用能力榜375
29.0
AA
数学推理242
36.0
AA
多模态榜69
60.0
LS
推理能力102
17.0
LS
科学能力329
33.0
AA

基準測試分數 (LLM Stats)

Biology

GPQA50.3%自報

Code

Aider-Polyglot9.8%自報
Aider-Polyglot Edit6.2%自報

Communication

Multi-IF57.2%自報
TAU-bench Retail22.6%自報
Multi-Challenge15.0%自報
TAU-bench Airline14.0%自報

Finance

MMLU80.1%自報

General

IFEval74.5%自報
MMMLU66.9%自報
MMMU55.4%自報
Internal API instruction following (hard)31.6%自報

Language

COLLIE42.5%自報

Long Context

OpenAI-MRCR: 2 needle 128k36.6%自報
OpenAI-MRCR: 2 needle 1M12.0%自報
ComplexFuncBench5.7%自報
Graphwalks parents >128k5.6%自報
Graphwalks BFS >128k2.9%自報

Math

MathVista56.2%自報
AIME 202429.4%自報

Multimodal

CharXiv-D73.9%自報
CharXiv-R40.5%自報

Reasoning

Graphwalks BFS <128k25.0%自報
Graphwalks parents <128k9.4%自報

AA 評測指數

Math Index
24.0
Intelligence Index
13.0
Coding Index
11.2
Math 500
0.8
Mmlu Pro
0.7
Gpqa
0.5
Livecodebench
0.3
Ifbench
0.3
Scicode
0.3
Aime 25
0.2
Aime
0.2
Tau2
0.2
Lcr
0.2
Hle
0.0
Terminalbench Hard
0.0

LLM Stats 分類評分

Finance
80
Legal
80
Healthcare
70
Instruction Following
70
Vision
60
Language
60
Math
60
Multimodal
60
Structured Output
50
Biology
50
Chemistry
50
General
50
Physics
50
Writing
40
Communication
30
Reasoning
30
Spatial Reasoning
10
Tool Calling
10
Code
10
Long Context
10

定價

輸入價格$0.1 / 1M tokens
輸出價格$0.4 / 1M tokens
混合價格(3:1)$0.175 / 1M tokens

速度

Tokens/秒121.7 tokens/s
首Token延遲0.52s
首回答延遲0.52s

可用提供商

(LS 內部計價單位)
提供商輸入價格輸出價格
OpenAI100K400K

外部連結