GLM-5.3-Flash
Z AIGLMОткрытые весаMIT · Коммерческое использование
Описание
GLM-5.3-Flash is the first natively multimodal model in Z.AI's GLM-5 series: a 320B-total / 18B-active MoE with hybrid sparse and linear attention (plus mHC), a 1M-token context window, and strong coding and long-horizon agent performance. MIT open weights are on Hugging Face. Launched 2026-08-26 after stealth testing as ox-alpha on OpenRouter and OpenCode. List pricing is $0.15/$0.50 per million input/output tokens (a temporary 50% promo ends 2026-09-09 16:00 UTC).
Дата выхода
2026-08-26
Параметры
320.0B
Длина контекста
1.0M
Модальности
image, pdf, text, video
Радар способностей
41
general
68
coding
91
reasoning
67
science
60
agents
80
multimodal
Рейтинги
| Домен | #Место | Оценка | Источник |
|---|---|---|---|
| Агентные возможности | 34 | 55.0 | LS |
| Рейтинг кодинга | 38 | 89.0 | AA |
| Общий рейтинг | 41 | 71.0 | AA |
| Математическое мышление | 49 | 81.0 | LB |
| Мультимодальный рейтинг | 67 | 54.0 | LS |
| Рассуждения | 41 | 78.0 | LB |
| Наука | 47 | 77.0 | AA |
Оценки бенчмарков (LLM Stats)
(LLM Stats (zeroeval))Agents
Toolathlon
78.4%Сам.
OfficeQA Pro
62.4%Сам.
AutomationBench
48.8%Сам.
Agents' Last Exam
26.3%Сам.
Code
DeepSWE 1.1
63.4%Сам.
NL2Repo
56.3%Сам.
General
GDPval-AA
1773.00 / 3000Сам.
Artificial Analysis
57.0%Сам.
Reasoning
CharXiv-R
89.4%Сам.
Terminal-Bench 2.1
84.3%Сам.
Humanity's Last Exam
55.3%Сам.
Vision
MMVU
80.5%Сам.
Chartography
78.0%Сам.
MVBench
77.8%Сам.
BabyVision
53.4%Сам.
Индексы оценки AA
(Artificial Analysis)Gpqa(NYU + Cohere + Anthropic (2023))91.2
Terminalbench V2 184.3
Lcr(Artificial Analysis)80.0
Coding Index(Artificial Analysis)71.5
Scicode(UIUC + Argonne National Lab (2024))51.6
Tau Banking47.2
Intelligence Index(Artificial Analysis)41.8
Hle(Center for AI Safety + Scale AI (2025))39.9
Terminalbench V4 032.8
Оценки категорий LLM Stats
(LLM Stats (zeroeval))Legal100
Finance100
Agents100
Reasoning100
General100
Multimodal80
Spatial Reasoning80
Video80
Code70
Vision70
Math60
Tool Calling60
Цены
Цена ввода$0.15 / 1M токенов
Цена вывода$0.5 / 1M токенов
Смешанная цена (3:1)$0.237 / 1M токенов
Цена чтения кэша$0.03 / 1M токенов
Цена записи кэшаБесплатно
Скорость
Токенов/сек62.8
Задержка первого токена2.46s
Время до первого ответа34.31s
Рейтинг цен провайдеров
Рейтинг цен провайдеров
48 провайдеров
Самый дешевый: ZAIСамый дорогой: Privatemode AI
ПровайдерВводВывод
1ZAIСамый дешевый
$0
$0
2DeepInfra
$0
$0
3Novita
$0
$0
4FriendliAI
$0
$0
5Pareto Inference
$0.06
$0.2
6302.AI
$0.075
$0.25
7ZenMux
$0.075
$0.25
8EmpirioLabs AI
$0.075
$0.25
9NEAR AI Cloud
$0.08
$0.5
10IteraCompute
$0.08
$0.28
11Standard Compute
$0.08
$0.5
12DevPass (LLM Gateway)
$0.088
$0.25
13NanoGPT
$0.1
$0.3
14Vultr
$0.1
$0.35
15Cortecs
$0.1
$0.35
16AIHubMix
$0.11268
$0.39438
17Kenari
$0.11429
$0.34286
18Melious
$0.11592
$0.46368
19Zhipu AI
$0.11952
$0.41831
20Vancine
$0.12
$0.4
21Boundless
$0.12
$0.4
22GreenPT
$0.12775
$0.51102
23engy
$0.135
$0.45
24Z AIОсновной
$0.15
$0.5
25OpenRouter
$0.15
$0.5
26OpenCode Go
$0.15
$0.5
27Vivgrid
$0.15
$0.5
28Kilo Gateway
$0.15
$0.5
29Z.AI
$0.15
$0.5
30OpenCode
$0.15
$0.5
31Ollama Cloud
$0.15
$0.5
32DigitalOcean
$0.15
$0.5
33Neuralwatt
$0.15
$0.5
34OrcaRouter
$0.15
$0.5
35Neon
$0.15
$0.5
36CrossModel
$0.15
$0.5
37Ofox
$0.15
$0.5
38TokenGo
$0.15
$0.5
39Prime Inference
$0.15
$0.5
40AI/ML API
$0.15
$0.5
41CommandCode
$0.15
$0.5
42Bothub
$0.1566
$0.522
43Charm Hyper
$0.16332
$0.5444
44above.dev
$0.165
$0.55
45Requesty
$0.2
$0.6
46TensorX
$0.2
$0.5
47Opper
$0.2
$0.5
48Privatemode AI
$0.2311
$0.7511
Сравнение цен разных API-провайдеров для этой модели.