跳转到主要内容
AITier

Qwen3.5 35B A3B (Non-reasoning)

AlibabaQwen开源权重Apache 2.0 · 商用许可

描述

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.

发布日期
2026-02-24
参数规模
35.0B
上下文长度
262K
支持模态
audio, image, text, video

能力雷达图

15
general
37
coding
82
reasoning
61
science
60
agents
80
multimodal

排行榜排名

领域#排名分数来源
智能体能力模型榜62
46.0
LS
代码能力榜339
44.0
AA
通用能力榜286
44.0
AA
多模态榜102
46.0
LS
科学能力247
52.0
AA

基准测试分数 (LLM Stats)

(LLM Stats (zeroeval))

Agents

t2-bench81.2%自报
VITA-Bench31.9%自报
DeepPlanning22.8%自报

Chat

IFEvalGoogle Research (2023)91.9%自报
Multi-Challenge60.0%自报

Code

FullStackBench en58.1%自报
FullStackBench zh55.0%自报

General

C-Eval90.2%自报
MAXIFE86.6%自报
Include79.7%自报
AndroidWorld_SR71.1%自报
NOVA-6357.1%自报

Healthcare

PMC-VQA62.0%自报
MedXpertQA61.4%自报

Instruction Following

IFBench70.2%自报

Language

MMLU-Redux93.3%自报
MMLU-Pro85.3%自报
MMMLU85.2%自报
MMLU-ProX81.0%自报
WMT24++76.3%自报

Long Context

LongBench v259.0%自报

Math

HMMT2589.2%自报
HMMT 202589.0%自报
MathVista-Mini86.2%自报
DynaMath85.0%自报
MathVision83.9%自报
CodeForces0.82 / 3000自报
PolyMATH64.4%自报

Multimodal

VideoMME w/o sub.82.5%自报
MMMU81.4%自报
VideoMMMU80.4%自报
TIR-Bench55.5%自报
OSWorld-Verified54.5%自报

Reasoning

Global PIQA86.6%自报
GPQANYU + Cohere + Anthropic (2023)84.2%自报
CharXiv-R77.5%自报
LiveCodeBench v674.6%自报
BrowseComp-zh69.5%自报
SWE-Bench Verified69.2%自报
SuperGPQA63.4%自报
BrowseCompOpenAI (2025)61.0%自报
AA-LCR58.5%自报
Humanity's Last Exam47.4%自报
Seal-041.4%自报
Terminal-Bench 2.0Stanford × Laude Institute (2026)40.5%自报
OJBench36.0%自报

Search

WideSearch57.1%自报

Tool Calling

BFCL-V467.3%自报

Video

MLVU85.6%自报

Vision

CountBench0.98 / 100自报
VLMsAreBlind97.0%自报
V*92.7%自报
AI2D92.6%自报
MMBench-V1.191.5%自报
OCRBench91.0%自报
OmniDocBench 1.589.3%自报
RefCOCO-avg0.89 / 100自报
VideoMME w sub.86.6%自报
RealWorldQA84.1%自报
EmbSpatialBench0.83 / 100自报
MMStar81.9%自报
CC-OCR80.7%自报
LingoQA79.2%自报
SlakeVQA78.7%自报
MMMU-Pro75.1%自报
MVBench74.8%自报
MMVU72.3%自报
LVBench71.4%自报
ScreenSpot Pro68.6%自报
Hallusion Bench67.9%自报
ERQA64.8%自报
RefSpatialBench0.64 / 100自报
MMLongBench-Doc0.59 / 100自报
SimpleVQA0.58 / 100自报
ODinW42.6%自报
BabyVision38.4%自报
ZEROBench-Sub0.34 / 100自报
SUNRGBD0.33 / 100自报
Nuscene14.6%自报
Hypersim0.13 / 100自报
ZEROBench0.08 / 100自报

AA 评测指数

(Artificial Analysis)
Tau2(Sierra + U Toronto + Vector Institute (2025))
86.3
Gpqa(NYU + Cohere + Anthropic (2023))
81.9
Lcr(Artificial Analysis)
63.0
Ifbench(Google Research (2023))
44.5
Terminalbench V2 1
40.8
Coding Index(Artificial Analysis)
37.0
Intelligence Index(Artificial Analysis)
15.1
Hle(Center for AI Safety + Scale AI (2025))
13.4
Terminalbench Hard(Stanford × Laude Institute (2026))
10.6
Tau Banking
4.9

LLM Stats 分类评分

(LLM Stats (zeroeval))
Chat
80
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

定价

输入价格$0.25 / 1M tokens
输出价格$2 / 1M tokens
混合价格(3:1)$0.688 / 1M tokens

速度

Tokens/秒0.0
首Token延迟0.00s
首回答延迟0.00s

供应商价格排行

供应商价格排行

6 个供应商

最便宜: DeepInfra最贵: DevPass (LLM Gateway)
供应商输入输出
1DeepInfra最便宜
$0
$0
2AIHubMix
$0.0564
$0.4512
3302.AI
$0.06
$0.46
4Requesty
$0.14
$1
5Alibaba主要
$0.25
$2
6DevPass (LLM Gateway)
$0.25
$2

比较该模型在不同 API 供应商之间的定价。

外部链接

LLM StatsArtificial Analysis