跳转到主要内容

GPT-5.4 nano (medium)

OpenAIGPTProprietary

描述

GPT-5.4 nano is OpenAI's cheapest GPT-5.4-class model for simple high-volume tasks. It is optimized for classification, data extraction, ranking, and lightweight coding subagents, while preserving strong reasoning, multimodal understanding, and tool use for latency-sensitive workloads.

发布日期
2026-03-17
参数规模
上下文长度
400K
支持模态
image, text

能力雷达图

28
general
38
coding
76
reasoning
51
science
60
agents
85
multimodal

排行榜排名

领域#排名分数来源
智能体能力模型榜147
22.0
LS
代码能力榜144
65.0
AA
通用能力榜196
56.0
AA
多模态榜58
43.0
LS
科学能力191
56.0
AA

基准测试分数 (LLM Stats)

(LLM Stats (zeroeval))

Agents

MCP Atlas56.1%自报
SWE-Bench ProPrinceton NLP (2024)52.4%自报
Terminal-Bench 2.0Stanford × Laude Institute (2026)46.3%自报
OSWorld-Verified39.0%自报
Finance Agent v238.2%
Toolathlon35.5%自报

Biology

GPQANYU + Cohere + Anthropic (2023)82.8%自报

Communication

Tau2 Telecom92.5%自报

General

LiveBench70.1%
MMMU-Pro66.1%自报
MRCR v2 (8-needle)33.1%自报

Math

Humanity's Last Exam24.3%自报

Multimodal

OmniDocBench 1.575.8%自报

Reasoning

Graphwalks BFS <128k73.4%自报
Graphwalks parents <128k50.8%自报

AA 评测指数

(Artificial Analysis)
Intelligence Index(Artificial Analysis)
30.8
Gpqa(NYU + Cohere + Anthropic (2023))
0.8
Lcr(Artificial Analysis)
0.7
Ifbench(Google Research (2023))
0.6
Tau2(Sierra + U Toronto + Vector Institute (2025))
0.5
Scicode(UIUC + Argonne National Lab (2024))
0.4
Terminalbench Hard(Stanford × Laude Institute (2026))
0.3
Hle(Center for AI Safety + Scale AI (2025))
0.2

LLM Stats 分类评分

(LLM Stats (zeroeval))
Spatial Reasoning
9
Vision
6
Reasoning
3
General
2
Communication
90
Physics
80
Structured Output
80
Biology
80
Chemistry
80
Multimodal
70
Tool Calling
60
Long Context
50
Math
50
Code
50
Finance
40
Agents
40
Healthcare
30

定价

输入价格$0.2 / 1M tokens
输出价格$1.25 / 1M tokens
混合价格(3:1)$0.463 / 1M tokens
缓存读取价格$0.02 / 1M tokens

速度

Tokens/秒0.0
首Token延迟0.00s
首回答延迟0.00s

供应商价格排行

供应商价格排行

1 个供应商

供应商输入输出
1OpenAI
$0
$0

比较该模型在不同 API 供应商之间的定价。

外部链接