跳转到主要内容
AITier

Qwen3.5 122B A10B (Non-reasoning)

AlibabaQwen开源权重Apache 2.0 · 商用许可

描述

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.

发布日期
2026-02-24
参数规模
122.0B
上下文长度
262K
支持模态
audio, image, text, video

能力雷达图

17
general
43
coding
83
reasoning
62
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47
49.0
LS
代码能力榜267
55.0
AA
通用能力榜245
48.0
AA
多模态榜43
59.0
LS
科学能力228
55.0
AA

基准测试分数 (LLM Stats)

(LLM Stats (zeroeval))

Agents

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

Chat

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

Code

FullStackBench en62.6%自报
FullStackBench zh58.7%自报

General

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

Healthcare

MedXpertQA67.3%自报
PMC-VQA63.3%自报

Instruction Following

IFBench76.1%自报

Language

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

Long Context

LongBench v260.2%自报

Math

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

Multimodal

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

Reasoning

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

Search

WideSearch60.5%自报

Tool Calling

BFCL-V472.2%自报

Video

MLVU87.3%自报

Vision

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

AA 评测指数

(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

LLM Stats 分类评分

(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

定价

输入价格$0.4 / 1M tokens
输出价格$3.2 / 1M tokens
混合价格(3:1)$1.1 / 1M tokens

速度

Tokens/秒142.8
首Token延迟1.03s
首回答延迟1.03s

供应商价格排行

供应商价格排行

2 个供应商

最便宜: DeepInfra最贵: Alibaba
供应商输入输出
1DeepInfra最便宜
$0
$0
2Alibaba主要
$0.4
$3.2

比较该模型在不同 API 供应商之间的定价。

外部链接

LLM StatsArtificial Analysis