o3
OpenAIOpenAI o-seriesProprietary
Описание
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.
Дата выхода
2025-04-16
Параметры
—
Длина контекста
200K
Модальности
file, image, text
Радар способностей
50
general
54
coding
90
reasoning
56
scienceоцен.
70
agents
85
multimodal
Science использует прокси на основе рассуждений, когда специализированные научные бенчмарки недоступны.
Рейтинги
| Домен | #Место | Оценка | Источник |
|---|---|---|---|
| Agents & Tools | 50 | 57.0 | LS |
| Code Ranking | 41 | 74.0 | AA |
| General Ranking | 55 | 77.0 | AA |
| Math Reasoning | 28 | 92.0 | AA |
| Multimodal Ranking | 33 | 79.0 | LS |
| Reasoning | 81 | 53.0 | LS |
| Science | 73 | 66.0 | AA |
Оценки бенчмарков (LLM Stats)
Agents
Tau-bench
63.0%Сам.
BrowseComp
49.7%Сам.
Biology
GPQA
83.3%Сам.
Code
Aider-Polyglot
81.3%Сам.
SWE-Bench Verified
69.1%Сам.
Communication
Tau2 Retail
80.2%Сам.
Tau2 Airline
64.8%Сам.
Multi-Challenge
60.4%Сам.
Tau2 Telecom
58.2%Сам.
General
MMMU
82.9%Сам.
MMMU-Pro
76.4%Сам.
Healthcare
VideoMMMU
83.3%Сам.
Language
COLLIE
98.4%Сам.
Math
AIME 2024
91.6%Сам.
MathVista
86.8%Сам.
AIME 2025
86.4%Сам.
FrontierMath
15.8%Сам.
Humanity's Last Exam
14.7%Сам.
Multimodal
CharXiv-R
78.6%Сам.
Reasoning
ARC-AGI
88.0%Сам.
ERQA
64.0%Сам.
ARC-AGI v2
6.5%Сам.
Индексы оценки AA
Math Index88.3
Intelligence Index38.4
Coding Index38.4
Math 5001.0
Aime0.9
Aime 250.9
Mmlu Pro0.9
Gpqa0.8
Livecodebench0.8
Tau20.8
Ifbench0.7
Lcr0.7
Scicode0.4
Terminalbench Hard0.4
Hle0.2
Оценки категорий LLM Stats
Writing100
Language100
Biology80
Chemistry80
Code80
General80
Healthcare80
Multimodal80
Physics80
Tool Calling70
Communication70
Frontend Development70
Reasoning70
Vision60
Agents60
Math60
Spatial Reasoning50
Search50
Цены
Цена ввода$2 / 1M tokens
Цена вывода$8 / 1M tokens
Смешанная цена (3:1)$3.5 / 1M tokens
Скорость
Токенов/сек92.1 tokens/s
Задержка первого токена9.54s
Время до первого ответа9.54s
Доступные провайдеры
(Внутренние единицы LS)Нет данных провайдеров