o3-mini
OpenAIOpenAI o-seriesProprietary
描述
A smaller variant of O3, expected to offer enhanced multimodal capabilities, improved reasoning, and more efficient resource utilization compared to previous models while maintaining strong performance on core tasks.
發布日期
2025-01-31
參數規模
—
上下文長度
200K
支援模態
file, text
能力雷達圖
39
general
39
coding
83
reasoning
49
science估算
40
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
基準測試分數 (LLM Stats)
Biology
GPQA
77.2%自報
Code
Aider-Polyglot
66.7%自報
Aider-Polyglot Edit
60.4%自報
SWE-Bench Verified
49.3%自報
SWE-Lancer
18.0%自報
SWE-Lancer (IC-Diamond subset)
7.4%自報
Communication
Multi-IF
79.5%自報
TAU-bench Retail
57.6%自報
Multi-Challenge
39.9%自報
TAU-bench Airline
32.4%自報
Factuality
SimpleQA
15.0%自報
Finance
MMLU
86.9%自報
General
IFEval
93.9%自報
LiveBench
84.6%自報
Multilingual MMLU
80.7%自報
Internal API instruction following (hard)
50.0%自報
Language
COLLIE
98.7%自報
Long Context
OpenAI-MRCR: 2 needle 128k
18.7%自報
ComplexFuncBench
17.6%自報
Math
MATH
97.9%自報
MGSM
92.0%自報
AIME 2024
87.3%自報
FrontierMath
9.2%自報
Reasoning
Graphwalks parents <128k
58.3%自報
Graphwalks BFS <128k
51.0%自報
AA 評測指數
Intelligence Index25.9
Coding Index17.9
Math 5001.0
Mmlu Pro0.8
Aime0.8
Gpqa0.7
Livecodebench0.7
Scicode0.4
Tau20.3
Hle0.1
Terminalbench Hard0.1
LLM Stats 分類評分
Writing100
Finance90
Healthcare90
Instruction Following90
Language90
Legal90
Biology80
Chemistry80
Math80
Physics80
General70
Structured Output60
Reasoning60
Spatial Reasoning50
Communication50
Frontend Development50
Tool Calling40
Code40
Long Context20
Factuality10
定價
輸入價格$1.1 / 1M tokens
輸出價格$4.4 / 1M tokens
混合價格(3:1)$1.925 / 1M tokens
速度
Tokens/秒135.1 tokens/s
首Token延遲10.07s
首回答延遲10.07s
可用提供商
(LS 內部計價單位)暫無提供商資料