Перейти к основному содержанию
AITier

Qwen3.5 27B (Reasoning)

AlibabaQwenОткрытые весаApache 2.0 · Коммерческое использование

Описание

Qwen3.5-27B is a multimodal dense foundation model with 27 billion parameters. It combines strong reasoning, coding, multilingual, long-context, and visual understanding performance in a production-friendly open-weight package with a native 262K context window.

Дата выхода
2026-02-24
Параметры
27.0B
Длина контекста
262K
Модальности
audio, image, text, video

Радар способностей

23
general
60
coding
86
reasoning
67
science
60
agents
80
multimodal

Рейтинги

Домен#МестоОценкаИсточник
Агентные возможности57
48.0
LS
Рейтинг кодинга183
69.0
AA
Общий рейтинг117
64.0
AA
Мультимодальный рейтинг74
53.0
LS
Наука163
63.0
AA

Оценки бенчмарков (LLM Stats)

(LLM Stats (zeroeval))

Agents

t2-bench79.0%Сам.
VITA-Bench41.9%Сам.
DeepPlanning22.6%Сам.

Chat

IFEvalGoogle Research (2023)95.0%Сам.
Multi-Challenge60.8%Сам.

Code

FullStackBench en60.1%Сам.
FullStackBench zh57.4%Сам.

General

C-Eval90.5%Сам.
MAXIFE88.0%Сам.
Include81.6%Сам.
AndroidWorld_SR64.2%Сам.
NOVA-6358.1%Сам.

Healthcare

MedXpertQA62.4%Сам.
PMC-VQA62.4%Сам.

Instruction Following

IFBench76.5%Сам.

Language

MMLU-Redux93.2%Сам.
MMLU-Pro86.1%Сам.
MMMLU85.9%Сам.
MMLU-ProX82.2%Сам.
WMT24++77.6%Сам.

Long Context

LongBench v260.6%Сам.

Math

HMMT 202592.0%Сам.
HMMT2589.8%Сам.
MathVista-Mini87.8%Сам.
DynaMath87.7%Сам.
MathVision86.0%Сам.
CodeForces0.81 / 3000Сам.
PolyMATH71.2%Сам.

Multimodal

VideoMME w/o sub.82.8%Сам.
VideoMMMU82.3%Сам.
MMMU82.3%Сам.
TIR-Bench59.8%Сам.
OSWorld-Verified56.2%Сам.

Reasoning

Global PIQA87.5%Сам.
GPQANYU + Cohere + Anthropic (2023)85.5%Сам.
LiveCodeBench v680.7%Сам.
CharXiv-R79.5%Сам.
SWE-Bench Verified72.4%Сам.
AA-LCR66.1%Сам.
SuperGPQA65.6%Сам.
BrowseComp-zh62.1%Сам.
BrowseCompOpenAI (2025)61.0%Сам.
Humanity's Last Exam48.5%Сам.
Seal-047.2%Сам.
Terminal-Bench 2.0Stanford × Laude Institute (2026)41.6%Сам.
OJBench40.1%Сам.

Search

WideSearch61.1%Сам.

Tool Calling

BFCL-V468.5%Сам.

Video

MLVU85.9%Сам.

Vision

CountBench0.98 / 100Сам.
VLMsAreBlind96.9%Сам.
V*93.7%Сам.
AI2D92.9%Сам.
MMBench-V1.192.6%Сам.
RefCOCO-avg0.91 / 100Сам.
OCRBench89.4%Сам.
OmniDocBench 1.588.9%Сам.
VideoMME w sub.87.0%Сам.
EmbSpatialBench0.84 / 100Сам.
RealWorldQA83.7%Сам.
LingoQA82.0%Сам.
CC-OCR81.0%Сам.
MMStar81.0%Сам.
SlakeVQA80.0%Сам.
MMMU-Pro75.0%Сам.
MVBench74.6%Сам.
LVBench73.6%Сам.
MMVU73.3%Сам.
ScreenSpot Pro70.3%Сам.
Hallusion Bench70.0%Сам.
RefSpatialBench0.68 / 100Сам.
ERQA60.5%Сам.
MMLongBench-Doc0.60 / 100Сам.
SimpleVQA0.56 / 100Сам.
BabyVision44.6%Сам.
ODinW41.1%Сам.
ZEROBench-Sub0.36 / 100Сам.
SUNRGBD0.35 / 100Сам.
Nuscene15.2%Сам.
Hypersim0.13 / 100Сам.
ZEROBench0.10 / 100Сам.

Индексы оценки AA

(Artificial Analysis)
Tau2(Sierra + U Toronto + Vector Institute (2025))
93.9
Gpqa(NYU + Cohere + Anthropic (2023))
85.8
Lcr(Artificial Analysis)
77.7
Ifbench(Google Research (2023))
75.6
Terminalbench Hard(Stanford × Laude Institute (2026))
32.6
Hle(Center for AI Safety + Scale AI (2025))
23.9
Intelligence Index(Artificial Analysis)
22.9

Оценки категорий LLM Stats

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

Цены

Цена ввода$0.3 / 1M токенов
Цена вывода$2.4 / 1M токенов
Смешанная цена (3:1)$0.825 / 1M токенов

Скорость

Токенов/сек0.0
Задержка первого токена0.00s
Время до первого ответа0.00s

Рейтинг цен провайдеров

Рейтинг цен провайдеров

4 провайдеров

Самый дешевый: DeepInfraСамый дорогой: Alibaba
ПровайдерВводВывод
1DeepInfraСамый дешевый
$0
$0
2Novita
$0
$0
3EmpirioLabs AI
$0.086
$0.688
4AlibabaОсновной
$0.3
$2.4

Сравнение цен разных API-провайдеров для этой модели.

Внешние ссылки

LLM StatsArtificial Analysis