Перейти к основному содержанию
AITier

Qwen3.5 122B A10B (Non-reasoning)

AlibabaQwenОткрытые весаApache 2.0 · Коммерческое использование

Описание

Qwen3.5-122B-A10B is a multimodal Mixture-of-Experts model with 122 billion total parameters and 10 billion activated parameters. It combines strong reasoning, coding, long-context, and visual understanding performance with production-friendly efficiency and a native 262K context window.

Дата выхода
2026-02-24
Параметры
122.0B
Длина контекста
262K
Модальности
audio, image, text, video

Радар способностей

17
general
43
coding
83
reasoning
62
science
60
agents
80
multimodal

Рейтинги

Домен#МестоОценкаИсточник
Агентные возможности47
49.0
LS
Рейтинг кодинга265
55.0
AA
Общий рейтинг244
48.0
AA
Мультимодальный рейтинг43
59.0
LS
Наука227
55.0
AA

Оценки бенчмарков (LLM Stats)

(LLM Stats (zeroeval))

Agents

t2-bench79.5%Сам.
VITA-Bench33.6%Сам.
DeepPlanning24.1%Сам.

Chat

IFEvalGoogle Research (2023)93.4%Сам.
Multi-Challenge61.5%Сам.

Code

FullStackBench en62.6%Сам.
FullStackBench zh58.7%Сам.

General

C-Eval91.9%Сам.
MAXIFE87.9%Сам.
Include82.8%Сам.
AndroidWorld_SR66.4%Сам.
NOVA-6358.6%Сам.

Healthcare

MedXpertQA67.3%Сам.
PMC-VQA63.3%Сам.

Instruction Following

IFBench76.1%Сам.

Language

MMLU-Redux94.0%Сам.
MMMLU86.7%Сам.
MMLU-Pro86.7%Сам.
MMLU-ProX82.2%Сам.
WMT24++78.3%Сам.

Long Context

LongBench v260.2%Сам.

Math

HMMT 202591.4%Сам.
HMMT2590.3%Сам.
MathVista-Mini87.4%Сам.
MathVision86.2%Сам.
DynaMath85.9%Сам.
CodeForces0.85 / 3000Сам.
PolyMATH68.9%Сам.

Multimodal

VideoMME w/o sub.83.9%Сам.
MMMU83.9%Сам.
VideoMMMU82.0%Сам.
OSWorld-Verified58.0%Сам.
TIR-Bench53.2%Сам.

Reasoning

Global PIQA88.4%Сам.
GPQANYU + Cohere + Anthropic (2023)86.6%Сам.
LiveCodeBench v678.9%Сам.
CharXiv-R77.2%Сам.
SWE-Bench Verified72.0%Сам.
BrowseComp-zh69.9%Сам.
SuperGPQA67.1%Сам.
AA-LCR66.9%Сам.
BrowseCompOpenAI (2025)63.8%Сам.
Terminal-Bench 2.0Stanford × Laude Institute (2026)49.4%Сам.
Humanity's Last Exam47.5%Сам.
Seal-044.1%Сам.
OJBench39.5%Сам.

Search

WideSearch60.5%Сам.

Tool Calling

BFCL-V472.2%Сам.

Video

MLVU87.3%Сам.

Vision

CountBench0.97 / 100Сам.
VLMsAreBlind96.7%Сам.
AI2D93.3%Сам.
V*93.2%Сам.
MMBench-V1.192.8%Сам.
OCRBench92.1%Сам.
RefCOCO-avg0.91 / 100Сам.
OmniDocBench 1.589.8%Сам.
VideoMME w sub.87.3%Сам.
RealWorldQA85.1%Сам.
EmbSpatialBench0.84 / 100Сам.
MMStar82.9%Сам.
CC-OCR81.8%Сам.
SlakeVQA81.6%Сам.
LingoQA80.8%Сам.
MMMU-Pro76.9%Сам.
MVBench76.6%Сам.
MMVU74.7%Сам.
LVBench74.4%Сам.
ScreenSpot Pro70.4%Сам.
RefSpatialBench0.69 / 100Сам.
Hallusion Bench67.6%Сам.
ERQA62.0%Сам.
SimpleVQA0.62 / 100Сам.
MMLongBench-Doc0.59 / 100Сам.
ODinW44.5%Сам.
BabyVision40.2%Сам.
SUNRGBD0.36 / 100Сам.
ZEROBench-Sub0.36 / 100Сам.
Nuscene15.4%Сам.
Hypersim0.13 / 100Сам.
ZEROBench0.09 / 100Сам.

Индексы оценки AA

(Artificial Analysis)
Tau2(Sierra + U Toronto + Vector Institute (2025))
84.5
Gpqa(NYU + Cohere + Anthropic (2023))
82.7
Lcr(Artificial Analysis)
61.3
Ifbench(Google Research (2023))
50.8
Terminalbench V2 1
47.2
Coding Index(Artificial Analysis)
43.3
Terminalbench Hard(Stanford × Laude Institute (2026))
29.5
Intelligence Index(Artificial Analysis)
17.7
Hle(Center for AI Safety + Scale AI (2025))
15.9
Tau Banking
10.3

Оценки категорий LLM Stats

(LLM Stats (zeroeval))
Biology
90
Chat
80
Image To Text
80
Instruction Following
80
Language
80
Legal
80
Math
80
Physics
80
Structured Output
80
Embodied
80
Finance
80
Grounding
80
Healthcare
80
Chemistry
80
Text-to-image
80
Video
80
Long Context
70
Multimodal
70
Reasoning
70
Spatial Reasoning
70
Frontend Development
70
General
70
Economics
70
Vision
70
Search
60
Agents
60
Code
60
Communication
60
Tool Calling
60
Spatial
20
3d
20

Цены

Цена ввода$0.4 / 1M токенов
Цена вывода$3.2 / 1M токенов
Смешанная цена (3:1)$1.1 / 1M токенов

Скорость

Токенов/сек145.3
Задержка первого токена1.01s
Время до первого ответа1.01s

Рейтинг цен провайдеров

Рейтинг цен провайдеров

2 провайдеров

Самый дешевый: DeepInfraСамый дорогой: Alibaba
ПровайдерВводВывод
1DeepInfraСамый дешевый
$0
$0
2AlibabaОсновной
$0.4
$3.2

Сравнение цен разных API-провайдеров для этой модели.

Внешние ссылки

LLM StatsArtificial Analysis