Qwen3.5 122B A10B (Non-reasoning)

AlibabaQwenOpen WeightApache 2.0 · Commercial OK

Description

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.

Release Date

2026-02-24

Parameters

122.0B

Context Length

262K

Modalities

audio, image, text, video

Capability Radar

general

coding

reasoning

science

agents

multimodal

Rankings

Domain	#Rank	Score	Source
Agentic Capability	47	49.0	LS
Code Ranking	268	55.0	AA
General Ranking	246	48.0	AA
Multimodal Ranking	43	59.0	LS
Science	229	55.0	AA

Benchmark Scores (LLM Stats)

(LLM Stats (zeroeval))

Agents

t2-bench

79.5%SR

VITA-Bench

33.6%SR

DeepPlanning

24.1%SR

Chat

IFEvalGoogle Research (2023)

93.4%SR

Multi-Challenge

61.5%SR

Code

FullStackBench en

62.6%SR

FullStackBench zh

58.7%SR

General

C-Eval

91.9%SR

MAXIFE

87.9%SR

Include

82.8%SR

AndroidWorld_SR

66.4%SR

NOVA-63

58.6%SR

Healthcare

MedXpertQA

67.3%SR

PMC-VQA

63.3%SR

Instruction Following

IFBench

76.1%SR

Language

MMLU-Redux

94.0%SR

MMMLU

86.7%SR

MMLU-Pro

86.7%SR

MMLU-ProX

82.2%SR

WMT24++

78.3%SR

Long Context

LongBench v2

60.2%SR

Math

HMMT 2025

91.4%SR

HMMT25

90.3%SR

MathVista-Mini

87.4%SR

MathVision

86.2%SR

DynaMath

85.9%SR

CodeForces

0.85 / 3000SR

PolyMATH

68.9%SR

Multimodal

VideoMME w/o sub.

83.9%SR

MMMU

83.9%SR

VideoMMMU

82.0%SR

OSWorld-Verified

58.0%SR

TIR-Bench

53.2%SR

Reasoning

Global PIQA

88.4%SR

GPQANYU + Cohere + Anthropic (2023)

86.6%SR

LiveCodeBench v6

78.9%SR

CharXiv-R

77.2%SR

SWE-Bench Verified

72.0%SR

BrowseComp-zh

69.9%SR

SuperGPQA

67.1%SR

AA-LCR

66.9%SR

BrowseCompOpenAI (2025)

63.8%SR

Terminal-Bench 2.0Stanford × Laude Institute (2026)

49.4%SR

Humanity's Last Exam

47.5%SR

Seal-0

44.1%SR

OJBench

39.5%SR

Search

WideSearch

60.5%SR

Tool Calling

BFCL-V4

72.2%SR

Video

MLVU

87.3%SR

Vision

CountBench

0.97 / 100SR

VLMsAreBlind

96.7%SR

AI2D

93.3%SR

93.2%SR

MMBench-V1.1

92.8%SR

OCRBench

92.1%SR

RefCOCO-avg

0.91 / 100SR

OmniDocBench 1.5

89.8%SR

VideoMME w sub.

87.3%SR

RealWorldQA

85.1%SR

EmbSpatialBench

0.84 / 100SR

MMStar

82.9%SR

CC-OCR

81.8%SR

SlakeVQA

81.6%SR

LingoQA

80.8%SR

MMMU-Pro

76.9%SR

MVBench

76.6%SR

MMVU

74.7%SR

LVBench

74.4%SR

ScreenSpot Pro

70.4%SR

RefSpatialBench

0.69 / 100SR

Hallusion Bench

67.6%SR

ERQA

62.0%SR

SimpleVQA

0.62 / 100SR

MMLongBench-Doc

0.59 / 100SR

ODinW

44.5%SR

BabyVision

40.2%SR

SUNRGBD

0.36 / 100SR

ZEROBench-Sub

0.36 / 100SR

Nuscene

15.4%SR

Hypersim

0.13 / 100SR

ZEROBench

0.09 / 100SR

AA Evaluation Indices

(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 Category Scores

(LLM Stats (zeroeval))

Biology

Chat

Image To Text

Instruction Following

Language

Legal

Math

Physics

Structured Output

Embodied

Finance

Grounding

Healthcare

Chemistry

Text-to-image

Video

Long Context

Multimodal

Reasoning

Spatial Reasoning

Frontend Development

General

Economics

Vision

Agents

Code

Communication

Tool Calling

Spatial

Pricing

Input Price$0.4 / 1M tokens

Output Price$3.2 / 1M tokens

Blended Price (3:1)$1.1 / 1M tokens

Speed

Tokens/sec143.3

Time to First Token1.09s

Time to Answer1.09s

Provider Price Ranking

2 providers

Cheapest: DeepInfraMost Expensive: Alibaba

ProviderInputOutput

1DeepInfraCheapest

2AlibabaPRIMARY

$0.4

$3.2

Compare pricing across different API providers for this model.

External Sources

LLM Stats Artificial Analysis