Перейти к основному содержанию
AITier

Qwen3.5 35B A3B (Non-reasoning)

AlibabaQwenОткрытые весаApache 2.0 · Коммерческое использование

Описание

Qwen3.5-35B-A3B is a multimodal Mixture-of-Experts model with 35 billion total parameters and 3 billion activated parameters. It combines strong reasoning, coding, agentic, and visual understanding performance with production-friendly efficiency and a native 262K context window.

Дата выхода
2026-02-24
Параметры
35.0B
Длина контекста
262K
Модальности
audio, image, text, video

Радар способностей

22
general
36
coding
82
reasoning
50
science
60
agents
80
multimodal

Рейтинги

Домен#МестоОценкаИсточник
Агентные возможности95
40.0
LS
Рейтинг кодинга248
46.0
AA
Общий рейтинг219
52.0
AA
Мультимодальный рейтинг44
47.0
LS
Наука218
52.0
AA

Оценки бенчмарков (LLM Stats)

(LLM Stats (zeroeval))

3d

SUNRGBD0.33 / 100Сам.
Hypersim0.13 / 100Сам.

Agents

t2-bench81.2%Сам.
AndroidWorld_SR71.1%Сам.
BFCL-V467.3%Сам.
BrowseCompOpenAI (2025)61.0%Сам.
FullStackBench en58.1%Сам.
WideSearch57.1%Сам.
TIR-Bench55.5%Сам.
FullStackBench zh55.0%Сам.
OSWorld-Verified54.5%Сам.
Terminal-Bench 2.0Stanford × Laude Institute (2026)40.5%Сам.
VITA-Bench31.9%Сам.
DeepPlanning22.8%Сам.

Biology

GPQANYU + Cohere + Anthropic (2023)84.2%Сам.

Chemistry

SuperGPQA63.4%Сам.

Code

SWE-Bench Verified69.2%Сам.

Communication

Multi-Challenge60.0%Сам.

Embodied

EmbSpatialBench0.83 / 100Сам.

Finance

MMLU-Pro85.3%Сам.
MMLU-ProX81.0%Сам.

General

MMLU-Redux93.3%Сам.
IFEvalGoogle Research (2023)91.9%Сам.
C-Eval90.2%Сам.
MAXIFE86.6%Сам.
Global PIQA86.6%Сам.
MMMLU85.2%Сам.
MMStar81.9%Сам.
MMMU81.4%Сам.
Include79.7%Сам.
MMMU-Pro75.1%Сам.
LiveCodeBench v674.6%Сам.
IFBench70.2%Сам.
LongBench v259.0%Сам.
SimpleVQA0.58 / 100Сам.
NOVA-6357.1%Сам.

Grounding

RefCOCO-avg0.89 / 100Сам.
ScreenSpot Pro68.6%Сам.
RefSpatialBench0.64 / 100Сам.

Healthcare

VideoMMMU80.4%Сам.
SlakeVQA78.7%Сам.
PMC-VQA62.0%Сам.
MedXpertQA61.4%Сам.

Image To Text

OCRBench91.0%Сам.

Language

LingoQA79.2%Сам.
WMT24++76.3%Сам.

Long Context

MLVU85.6%Сам.
LVBench71.4%Сам.
MMLongBench-Doc0.59 / 100Сам.
AA-LCR58.5%Сам.

Math

HMMT2589.2%Сам.
HMMT 202589.0%Сам.
MathVista-Mini86.2%Сам.
DynaMath85.0%Сам.
MathVision83.9%Сам.
CodeForces0.82 / 3000Сам.
PolyMATH64.4%Сам.
Humanity's Last Exam47.4%Сам.

Multimodal

VLMsAreBlind97.0%Сам.
V*92.7%Сам.
AI2D92.6%Сам.
MMBench-V1.191.5%Сам.
OmniDocBench 1.589.3%Сам.
VideoMME w sub.86.6%Сам.
VideoMME w/o sub.82.5%Сам.
CC-OCR80.7%Сам.
CharXiv-R77.5%Сам.
MVBench74.8%Сам.
MMVU72.3%Сам.
BabyVision38.4%Сам.
ZEROBench-Sub0.34 / 100Сам.
Nuscene14.6%Сам.
ZEROBench0.08 / 100Сам.

Reasoning

CountBench0.98 / 100Сам.
BrowseComp-zh69.5%Сам.
Hallusion Bench67.9%Сам.
ERQA64.8%Сам.
Seal-041.4%Сам.
OJBench36.0%Сам.

Spatial Reasoning

RealWorldQA84.1%Сам.

Vision

ODinW42.6%Сам.

Индексы оценки AA

(Artificial Analysis)
Coding Index(Artificial Analysis)
37.0
Intelligence Index(Artificial Analysis)
24.3
Tau2(Sierra + U Toronto + Vector Institute (2025))
0.9
Gpqa(NYU + Cohere + Anthropic (2023))
0.8
Lcr(Artificial Analysis)
0.6
Ifbench(Google Research (2023))
0.4
Terminalbench V2 1
0.4
Scicode(UIUC + Argonne National Lab (2024))
0.3
Hle(Center for AI Safety + Scale AI (2025))
0.1
Terminalbench Hard(Stanford × Laude Institute (2026))
0.1
Tau Banking
0.0

Оценки категорий LLM Stats

(LLM Stats (zeroeval))
Image To Text
80
Instruction Following
80
Language
80
Legal
80
Math
80
Physics
80
Structured Output
80
Embodied
80
Finance
80
Biology
80
Text-to-image
80
Video
80
Long Context
70
Multimodal
70
Reasoning
70
Spatial Reasoning
70
Frontend Development
70
General
70
Grounding
70
Healthcare
70
Chemistry
70
Vision
70
Search
60
Code
60
Communication
60
Economics
60
Tool Calling
60
Agents
50
3d
20
Spatial
10

Цены

Цена ввода$0.25 / 1M токенов
Цена вывода$2 / 1M токенов
Смешанная цена (3:1)$0.688 / 1M токенов

Скорость

Токенов/сек164.3
Задержка первого токена1.22s
Время до первого ответа1.22s

Рейтинг цен провайдеров

Рейтинг цен провайдеров

1 провайдеров

ПровайдерВводВывод
1AlibabaОсновной
$0.25
$2

Сравнение цен разных API-провайдеров для этой модели.

Внешние ссылки

LLM StatsArtificial Analysis