o3
OpenAIOpenAI o-seriesProprietary
描述
OpenAI's most powerful reasoning model. o3 is a well-rounded and powerful model across domains. It sets a new standard for math, science, coding, and visual reasoning tasks. It also excels at technical writing and instruction-following. Use it to think through multi-step problems that involve analysis across text, code, and images.
发布日期
2025-04-16
参数规模
—
上下文长度
200K
支持模态
image, pdf, text
能力雷达图
46
general
72
coding
90
reasoning
56
science
70
agents
85
multimodal
排行榜排名
基准测试分数 (LLM Stats)
(LLM Stats (zeroeval))Agents
Tau-bench
63.0%自报
BrowseCompOpenAI (2025)
49.7%自报
Biology
GPQANYU + Cohere + Anthropic (2023)
83.3%自报
Code
Aider-Polyglot
81.3%自报
SWE-Bench Verified
69.1%自报
Communication
Tau2 Retail
80.2%自报
Tau2 Airline
64.8%自报
Multi-Challenge
60.4%自报
Tau2 Telecom
58.2%自报
General
MMMU
82.9%自报
MMMU-Pro
76.4%自报
Healthcare
VideoMMMU
83.3%自报
Language
COLLIE
98.4%自报
Math
AIME 2024
91.6%自报
MathVista
86.8%自报
AIME 2025
86.4%自报
FrontierMath
15.8%自报
Humanity's Last Exam
14.7%自报
Multimodal
CharXiv-R
78.6%自报
Reasoning
ARC-AGI
88.0%自报
ERQA
64.0%自报
ARC-AGI v2
6.5%
AA 评测指数
(Artificial Analysis)Math Index(Artificial Analysis)88.3
Intelligence Index(Artificial Analysis)31.1
Math 500(OpenAI (2024), subset of Hendrycks et al. MATH (2021))1.0
Aime(MAA (Mathematical Association of America))0.9
Aime 25(MAA (Mathematical Association of America))0.9
Mmlu Pro(TIGER-Lab (Univ. of Waterloo, Toronto, CMU, 2024))0.9
Gpqa(NYU + Cohere + Anthropic (2023))0.8
Livecodebench(UC Berkeley + MIT + Cornell (2024))0.8
Tau2(Sierra + U Toronto + Vector Institute (2025))0.8
Lcr(Artificial Analysis)0.7
Ifbench(Google Research (2023))0.7
Scicode(UIUC + Argonne National Lab (2024))0.4
Terminalbench Hard(Stanford × Laude Institute (2026))0.4
Hle(Center for AI Safety + Scale AI (2025))0.2
LLM Stats 分类评分
(LLM Stats (zeroeval))Language100
Writing100
Multimodal80
Physics80
Healthcare80
Biology80
Chemistry80
Code80
Reasoning70
Frontend Development70
General70
Communication70
Tool Calling70
Math60
Agents60
Vision60
Search50
Spatial Reasoning50
定价
输入价格$2 / 1M tokens
输出价格$8 / 1M tokens
混合价格(3:1)$3.5 / 1M tokens
缓存读取价格$0.5 / 1M tokens
速度
Tokens/秒118.5
首Token延迟8.10s
首回答延迟8.10s
供应商价格排行
供应商价格排行
16 个供应商
最便宜: Poe最贵: Jiekou.AI
供应商输入输出
1Poe最便宜
$1.8
$7.2
2OpenAI主要
$2
$8
3NanoGPT
$2
$8
4Abacus
$2
$8
5OpenRouter
$2
$8
6Kilo Gateway
$2
$8
7Cloudflare AI Gateway
$2
$8
8Helicone
$2
$8
9Azure Cognitive Services
$2
$8
10Vercel AI Gateway
$2
$8
11LLM Gateway
$2
$8
12Azure
$2
$8
13NEAR AI Cloud
$2
$8
14Merge Gateway
$2
$8
15Impossibl
$2
$8
16Jiekou.AI
$10
$40
比较该模型在不同 API 供应商之间的定价。