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.
發布日期
2025-04-16
參數規模
—
上下文長度
200K
支援模態
image, pdf, text
能力雷達圖
40
general
81
coding
90
reasoning
63
science
70
agents
85
multimodal
排行榜排名
基準測試分數 (LLM Stats)
(LLM Stats (zeroeval))Chat
Tau2 Airline
64.8%自報
Multi-Challenge
60.4%自報
Communication
Tau2 Retail
80.2%自報
Tau2 Telecom
58.2%自報
General
Aider-Polyglot
81.3%自報
Tau-bench
63.0%自報
Language
COLLIE
98.4%自報
Math
AIME 2024
91.6%自報
MathVista
86.8%自報
AIME 2025
86.4%自報
FrontierMath
15.8%自報
Multimodal
VideoMMMU
83.3%自報
MMMU
82.9%自報
Reasoning
ARC-AGI
88.0%自報
GPQANYU + Cohere + Anthropic (2023)
83.3%自報
CharXiv-R
78.6%自報
SWE-Bench Verified
69.1%自報
BrowseCompOpenAI (2025)
49.7%自報
Humanity's Last Exam
14.7%自報
ARC-AGI v2
6.5%
Vision
MMMU-Pro
76.4%自報
ERQA
64.0%自報
AA 評測指數
(Artificial Analysis)Math 500(OpenAI (2024), subset of Hendrycks et al. MATH (2021))99.2
Aime(MAA (Mathematical Association of America))90.3
Aime 25(MAA (Mathematical Association of America))88.3
Math Index(Artificial Analysis)88.3
Mmlu Pro(TIGER-Lab (Univ. of Waterloo, Toronto, CMU, 2024))85.3
Gpqa(NYU + Cohere + Anthropic (2023))82.7
Livecodebench(UC Berkeley + MIT + Cornell (2024))80.8
Tau2(Sierra + U Toronto + Vector Institute (2025))80.7
Lcr(Artificial Analysis)74.7
Ifbench(Google Research (2023))71.4
Terminalbench Hard(Stanford × Laude Institute (2026))37.1
Intelligence Index(Artificial Analysis)20.2
Hle(Center for AI Safety + Scale AI (2025))20.1
LLM Stats 分類評分
(LLM Stats (zeroeval))Language100
Writing100
Multimodal80
Physics80
Healthcare80
Biology80
Chemistry80
Code80
Reasoning70
Frontend Development70
General70
Communication70
Tool Calling70
Chat60
Math60
Agents60
Vision60
Search50
Spatial Reasoning50
定價
輸入價格$2 / 1M tokens
輸出價格$8 / 1M tokens
混合價格(3:1)$3.5 / 1M tokens
快取讀取價格$0.5 / 1M tokens
速度
Tokens/秒151.5
首Token延遲5.57s
首回答延遲5.57s
供應商價格排行
供應商價格排行
20 個供應商
最便宜: Poe最貴: Jiekou.AI
供應商輸入輸出
1Poe最便宜
$1.8
$7.2
2OpenAI主要
$2
$8
3302.AI
$2
$8
4NanoGPT
$2
$8
5Abacus
$2
$8
6OpenRouter
$2
$8
7Kilo Gateway
$2
$8
8Cloudflare AI Gateway
$2
$8
9Helicone
$2
$8
10AIHubMix
$2
$8
11Azure Cognitive Services
$2
$8
12Vercel AI Gateway
$2
$8
13DevPass (LLM Gateway)
$2
$8
14Azure
$2
$8
15NEAR AI Cloud
$2
$8
16Merge Gateway
$2
$8
17Impossibl
$2
$8
18Eden AI
$2
$8
19LLM Gateway
$2
$8
20Jiekou.AI
$10
$40
比較該模型在不同 API 供應商之間的定價。