跳转到主要内容
AITier

GPT-4o (Aug '24)

OpenAIGPTProprietary

描述

GPT-4o ('o' for 'omni') is a multimodal AI model that accepts text, audio, image, and video inputs, and generates text, audio, and image outputs. It matches GPT-4 Turbo performance on text and code, with improvements in non-English languages, vision, and audio understanding.

发布日期
2024-08-06
参数规模
—
上下文长度
128K
支持模态
image, text

能力雷达图

7
general
32
coding
40
reasoning
37
science
50
agents
90
multimodal

排行榜排名

领域#排名分数来源
代码能力榜432
31.0
AA
通用能力榜580
20.0
AA
多模态榜116
39.0
LS
科学能力510
26.0
AA

基准测试分数 (LLM Stats)

(LLM Stats (zeroeval))

Chat

IFEvalGoogle Research (2023)81.0%自报
Multi-IF60.9%自报
TAU-bench Retail60.3%自报
Tau2 Airline45.5%自报
Multi-Challenge40.3%自报

Communication

Tau2 Retail63.4%自报
Tau2 Telecom23.5%自报

Factuality

SimpleQA38.2%自报

General

MMLU85.7%自报
Aider-Polyglot30.7%自报
Internal API instruction following (hard)29.2%自报
Aider-Polyglot Edit18.2%自报

Language

MMMLU81.4%自报
MMLU-Pro74.7%自报
COLLIE61.0%自报

Long Context

ComplexFuncBench66.5%自报
OpenAI-MRCR: 2 needle 128k31.9%自报

Math

MathVista61.4%自报
AIME 202413.1%自报

Multimodal

MMMU72.2%自报
VideoMMMU61.2%自报

Reasoning

ChartQAMasry et al. (2022)85.7%自报
CharXiv-D85.3%自报
GPQANYU + Cohere + Anthropic (2023)70.1%自报
CharXiv-R58.8%自报
TAU-bench Airline42.8%自报
Graphwalks BFS <128k41.7%自报
Graphwalks parents <128k35.4%自报
SWE-Bench Verified33.2%自报
SWE-Lancer32.6%自报
SWE-Lancer (IC-Diamond subset)12.4%自报
Humanity's Last Exam5.3%自报

Vision

AI2D94.2%自报
DocVQADocVQA (2020)92.8%自报
EgoSchema72.2%自报
ActivityNet61.9%自报
MMMU-Pro59.9%自报
ERQA35.2%自报

AA 评测指数

(Artificial Analysis)
Math 500(OpenAI (2024), subset of Hendrycks et al. MATH (2021))
79.5
Gpqa(NYU + Cohere + Anthropic (2023))
52.1
Lcr(Artificial Analysis)
41.0
Ifbench(Google Research (2023))
36.0
Livecodebench(UC Berkeley + MIT + Cornell (2024))
31.7
Tau2(Sierra + U Toronto + Vector Institute (2025))
28.9
Aime(MAA (Mathematical Association of America))
11.7
Terminalbench Hard(Stanford × Laude Institute (2026))
8.3
Intelligence Index(Artificial Analysis)
7.7
Hle(Center for AI Safety + Scale AI (2025))
2.3

LLM Stats 分类评分

(LLM Stats (zeroeval))
Image To Text
90
Legal
80
Finance
80
Instruction Following
70
Language
70
Multimodal
70
Physics
70
Healthcare
70
Biology
70
Chemistry
70
Vision
70
Chat
60
Long Context
60
Structured Output
60
Writing
60
Math
50
Reasoning
50
General
50
Communication
50
Tool Calling
50
Spatial Reasoning
40
Factuality
40
Frontend Development
30
Code
30

定价

输入价格$2.5 / 1M tokens
输出价格$10 / 1M tokens
混合价格(3:1)$4.375 / 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

2 个供应商

最便宜: OpenAI最贵: Azure
供应商输入输出
1OpenAI最便宜
$0
$0.00001
2Azure
$0
$0.00001

比较该模型在不同 API 供应商之间的定价。

外部链接

LLM StatsArtificial Analysis