GPT-5.6 Luna (Max)
OpenAIGPTProprietary
Описание
GPT-5.6 Luna is the fastest and most cost-efficient tier of OpenAI's GPT-5.6 family, designed for cost-sensitive, high-volume workloads. It brings strong capability at OpenAI's lowest GPT-5.6 price, supports max reasoning effort, and has a 1.05M-token context window.
Дата выхода
2026-07-09
Параметры
—
Длина контекста
1.1M
Модальности
image, pdf, text
Радар способностей
38
general
69
coding
91
reasoning
68
science
60
agents
85
multimodal
Рейтинги
| Домен | #Место | Оценка | Источник |
|---|---|---|---|
| Агентные возможности | 111 | 33.0 | LS |
| Рейтинг кодинга | 29 | 90.0 | AA |
| Общий рейтинг | 81 | 63.0 | AA |
| Математическое мышление | 37 | 87.0 | LB |
| Мультимодальный рейтинг | 99 | 46.0 | LS |
| Рассуждения | 28 | 86.0 | LB |
| Наука | 37 | 79.0 | AA |
Оценки бенчмарков (LLM Stats)
(LLM Stats (zeroeval))Agents
Connectors
99.9%Сам.
Search and Function-Calling
89.7%Сам.
Capture-the-Flag Challenges (Internal)
85.2%Сам.
Toolathlon
53.4%Сам.
LifeSciBench
51.2%Сам.
Internal Research Debugging Evaluation
50.8%Сам.
Agents' Last Exam
50.3%Сам.
SEC-bench Pro
48.9%Сам.
RSI Index
41.9%Сам.
Big Finance Bench
36.0%Сам.
Management Consulting Tasks (Internal)
35.4%Сам.
PostTrainBench Lite
29.6%Сам.
KernelGen 1P
22.4%Сам.
AutomationBench
14.9%Сам.
ExploitGym
12.4%Сам.
GeneBench-Pro
10.8%Сам.
NanoGPT
1.7%Сам.
Code
BenchCAD (with Python tool)
73.9%Сам.
DeepSWE
67.2%Сам.
DeepSWE 1.1
67.0%
BenchCAD
63.1%Сам.
ExploitBench
33.2%Сам.
General
Artificial Analysis
51.0%
GDP.pdf
22.7%Сам.
Healthcare
HealthBench Consensus
95.1%Сам.
HealthBench
55.8%Сам.
HealthBench Professional
55.7%Сам.
HealthBench Hard
32.0%Сам.
Long Context
MRCR v2 (8-needle, 512K-1M)
41.3%Сам.
MRCR v2 (8-needle)
41.3%Сам.
Math
FrontierMath
78.6%Сам.
FrontierMath Tier 4 (v2)
58.5%Сам.
Multimodal
OSWorld 2.0
45.6%Сам.
Reasoning
GPQANYU + Cohere + Anthropic (2023)
92.3%Сам.
Terminal-Bench 2.1
84.7%Сам.
BrowseCompOpenAI (2025)
83.3%Сам.
Graphwalks BFS >128k
81.3%Сам.
SWE-Bench ProPrinceton NLP (2024)
62.7%Сам.
Graphwalks BFS 1M
51.2%Сам.
FrontierCode 1.1
39.8%
MedChemBench (Internal)
30.4%Сам.
Terminal-Bench 4.0
17.3%
ARC-AGI-3
0.2%Сам.
Vision
MMMU-Pro (with tools)
79.5%Сам.
MMMU-Pro
78.4%Сам.
Индексы оценки AA
(Artificial Analysis)Gpqa(NYU + Cohere + Anthropic (2023))91.1
Lcr(Artificial Analysis)83.7
Terminalbench V2 180.9
Coding Index(Artificial Analysis)71.4
Scicode(UIUC + Argonne National Lab (2024))53.6
Hle(Center for AI Safety + Scale AI (2025))39.5
Intelligence Index(Artificial Analysis)37.3
Tau Banking31.1
Terminalbench V4 011.6
Оценки категорий LLM Stats
(LLM Stats (zeroeval))Physics90
Biology90
Chemistry90
Search80
Math70
Spatial Reasoning70
Multimodal60
Tool Calling60
Vision60
Long Context50
Reasoning50
General50
Healthcare50
Agents50
Code50
Safety40
Finance40
Science30
Systems20
Цены
Цена ввода$0.2 / 1M токенов
Цена вывода$1.2 / 1M токенов
Смешанная цена (3:1)$0.45 / 1M токенов
Цена чтения кэша$0.02 / 1M токенов
Цена записи кэша$0.25 / 1M токенов
Скорость
Токенов/сек0.0
Задержка первого токена0.00s
Время до первого ответа0.00s
Рейтинг цен провайдеров
Рейтинг цен провайдеров
2 провайдеров
Самый дешевый: OpenAIСамый дорогой: Neon
ПровайдерВводВывод
1OpenAIСамый дешевый
$0
$0
2Neon
$0.2
$1.2
Сравнение цен разных API-провайдеров для этой модели.