llmboard.ai
Benchmarks
CompareRankings
llmboard.ai
Benchmarks
CompareRankings
HomeModelsQwen3.5 397B A17B

Alibaba Cloud / Qwen Team model product

Qwen3.5 397B A17B

Qwen3.5-397B-A17B is Qwen's flagship Mixture-of-Experts model with 397 billion total parameters and 17 billion activated parameters. It delivers state-of-the-art performance across knowledge, reasoning, coding, mathematics, multilingual understanding, instruction following, long context, and agent tasks.

Updated Aug 10, 2026. Default version: Qwen3.5-397B-A17B

Compare
LLMBoard score75.2Qwen3.5-397B-A17B
Coverage100%21 benchmark families
Context window262.1KTokens
Official input price$0.43Alibaba (China) API

On this page

  • Specification
  • Capability
  • Benchmarks
  • Arena
  • Pricing
  • Versions
  • About
  • Compare
  • Similar models
  • FAQ

Model specification

Structured fields from the published default version.

Version
Qwen3.5-397B-A17B
Released
Feb 16, 2026
Knowledge cutoff
Unknown
Parameters
397B
Context window
262.1K
Max output
65.5K
Inputs
audio, image, text, video
Outputs
text
Open weights
Yes
License
Apache 2.0

Capability profile

This profile uses the latest version under this unique model that has a calculated LLMBoard score. Arena and price are excluded.

Qwen3.5-397B-A17B category scores

Benchmark results

Published benchmark records for the scored version Qwen3.5-397B-A17B.

38 rows
Columns

Show columns

BrowseComp-zh0.7113100.0%CAug 7, 2026
NOVA-630.6111100.0%CAug 7, 2026
SecCodeBench0.711100.0%CAug 7, 2026
VITA-Bench0.5110100.0%CAug 7, 2026
BFCL-V40.721391.7%CAug 7, 2026
C-Eval0.921894.1%CAug 7, 2026
LongBench v20.621793.8%CAug 7, 2026
MMLU-Redux0.924897.9%CAug 7, 2026
DeepPlanning0.33975.0%CAug 7, 2026
HMMT250.932591.7%CAug 7, 2026
Include0.933193.3%CAug 7, 2026
MAXIFE0.931180.0%CAug 7, 2026
MMLU-ProX0.833293.5%CAug 7, 2026
AA-LCR0.741578.6%CAug 7, 2026
PolyMATH0.742386.4%CAug 7, 2026
Seal-00.54640.0%CAug 7, 2026
t2-bench0.942386.4%CAug 7, 2026
Global PIQA0.951366.7%CAug 7, 2026
MMLU-Pro0.9512996.9%CAug 7, 2026
Multi-Challenge0.752985.7%CAug 7, 2026
SuperGPQA0.753487.9%CAug 7, 2026
WideSearch0.75950.0%CAug 7, 2026
MCP-Mark0.56828.6%CAug 7, 2026
HMMT 20250.973381.3%CAug 7, 2026
IFBench0.882874.1%CAug 7, 2026
IFEval0.986589.1%CAug 7, 2026
WMT24++0.882368.2%CAug 7, 2026
AIME 20260.9101743.8%CAug 7, 2026
LiveCodeBench v60.8125378.8%CAug 7, 2026
IMO-AnswerBench0.8141927.8%CAug 7, 2026
MMMLU0.9174966.7%CAug 7, 2026
SWE-bench Multilingual0.7203442.4%CAug 7, 2026
Toolathlon0.4263116.7%CAug 7, 2026
BrowseComp0.7295850.9%CAug 7, 2026
GPQA0.92923387.9%CAug 7, 2026
SWE-Bench Verified0.82910472.8%CAug 7, 2026
Terminal-Bench 2.00.5324935.4%CAug 7, 2026
Humanity's Last Exam0.3459251.6%CAug 7, 2026

Arena results

Preference and agent-evaluation signals from published Arena datasets.

100 rows
Columns

Show columns

visionentity recognition111277.8180N/AAug 6, 2026
text factualityjapanese131446.8396N/AAug 6, 2026
vision style controlentity recognition171226.3180N/AAug 6, 2026
text factualitygerman231416.5465N/AAug 6, 2026
text factualitykorean251392.5543N/AAug 6, 2026
visioncreative writing vision251275.71,403N/AAug 6, 2026
text factualityindustry mathematical261460.52,625N/AAug 6, 2026
textjapanese281428.5605N/AAug 6, 2026
text factualitychinese281506.12,590N/AAug 6, 2026
text factualitypolish281427.7776N/AAug 6, 2026
visionenglish281272.910,086N/AAug 6, 2026
vision style controlcreative writing vision281258.61,403N/AAug 6, 2026
text factualitymath291456.92,622N/AAug 6, 2026
visionhomework301287.83,064N/AAug 6, 2026
visionchinese311321.41,179N/AAug 6, 2026
vision style controlhumor311246.7811N/AAug 6, 2026
text factualityfrench321476.91,552N/AAug 6, 2026
visionhumor321265.1811N/AAug 6, 2026
vision style controlchinese321298.31,179N/AAug 6, 2026
visionoverall341265.023,868N/AAug 6, 2026
visiondiagram341280.36,340N/AAug 6, 2026
vision style controlhomework341282.53,064N/AAug 6, 2026
visionocr351272.616,855N/AAug 6, 2026
vision style controlenglish351254.010,086N/AAug 6, 2026
vision style controldiagram391275.96,340N/AAug 6, 2026
vision style controlocr391261.916,855N/AAug 6, 2026
text factualityexpert401483.85,002N/AAug 6, 2026
textchinese411500.83,394N/AAug 6, 2026
text factualityspanish411437.81,378N/AAug 6, 2026
textpolish421445.21,275N/AAug 6, 2026
text factualityindustry life and physical and social science421476.68,283N/AAug 6, 2026
text factualityindustry medicine and healthcare421481.23,590N/AAug 6, 2026
text style controljapanese421429.1605N/AAug 6, 2026
vision style controloverall421247.523,868N/AAug 6, 2026
textgerman431447.8999N/AAug 6, 2026
text style controlchinese441500.43,394N/AAug 6, 2026
textindustry mathematical451458.93,500N/AAug 6, 2026
textfrench471460.12,097N/AAug 6, 2026
textindustry life and physical and social science471458.410,304N/AAug 6, 2026
textmath471451.53,622N/AAug 6, 2026
textspanish471450.01,950N/AAug 6, 2026
text factualityindustry legal and government501462.43,595N/AAug 6, 2026
text factualitylonger query501463.624,604N/AAug 6, 2026
text style controlindustry mathematical501460.33,500N/AAug 6, 2026
textexpert511464.96,272N/AAug 6, 2026
textindustry business and management and financial operations531432.512,916N/AAug 6, 2026
textindustry medicine and healthcare531453.64,711N/AAug 6, 2026
textindustry software and it services531463.225,731N/AAug 6, 2026
text factualitymulti turn531460.38,680N/AAug 6, 2026
text factualityrussian531452.75,018N/AAug 6, 2026
text factualitycoding541498.216,334N/AAug 6, 2026
text factualityindustry software and it services541483.824,660N/AAug 6, 2026
textlonger query551442.926,491N/AAug 6, 2026
text factualityhard prompts english551478.318,313N/AAug 6, 2026
textnon english561431.233,724N/AAug 6, 2026
textcoding571465.318,329N/AAug 6, 2026
text style controlexpert581479.06,272N/AAug 6, 2026
text style controlspanish581448.01,950N/AAug 6, 2026
webdevwebdev-react581388.713,509N/AAug 6, 2026
textmulti turn591445.710,769N/AAug 6, 2026
text factualityindustry business and management and financial operations591449.810,692N/AAug 6, 2026
text factualityindustry entertainment and sports and media591417.511,458N/AAug 6, 2026
text factualityinstruction following591441.219,129N/AAug 6, 2026
text style controlgerman591445.3999N/AAug 6, 2026
text style controlmath591449.13,622N/AAug 6, 2026
webdevoverall591400.016,511N/AAug 6, 2026
webdevwebdev591400.016,511N/AAug 6, 2026
textexclude ties601434.546,583N/AAug 6, 2026
texthard prompts601448.541,548N/AAug 6, 2026
text factualityindustry writing and literature and language601428.112,910N/AAug 6, 2026
text factualitynon english601435.133,524N/AAug 6, 2026
textoverall611438.563,719N/AAug 6, 2026
textinstruction following611422.520,914N/AAug 6, 2026
text factualityhard prompts611468.141,317N/AAug 6, 2026
webdevwebdev-html611412.22,017N/AAug 6, 2026
text factualityoverall621445.463,388N/AAug 6, 2026
texthard prompts english631451.420,256N/AAug 6, 2026
textrussian631431.46,539N/AAug 6, 2026
text factualityexclude ties631445.846,350N/AAug 6, 2026
text style controlfrench641467.12,097N/AAug 6, 2026
textindustry writing and literature and language651416.114,873N/AAug 6, 2026
text style controllonger query651453.926,491N/AAug 6, 2026
textenglish671443.829,993N/AAug 6, 2026
text factualityenglish671453.429,737N/AAug 6, 2026
text style controlindustry life and physical and social science671465.810,304N/AAug 6, 2026
text style controlnon english671433.333,724N/AAug 6, 2026
text style controlpolish681441.91,275N/AAug 6, 2026
textkorean691386.81,046N/AAug 6, 2026
text factualitycreative writing691414.28,165N/AAug 6, 2026
text style controlcoding691491.718,329N/AAug 6, 2026
text style controlmulti turn691452.210,769N/AAug 6, 2026
text style controlhard prompts701463.441,548N/AAug 6, 2026
text style controlhard prompts english721467.420,256N/AAug 6, 2026
text style controlindustry software and it services721480.525,731N/AAug 6, 2026
text style controlinstruction following721432.720,914N/AAug 6, 2026
textindustry legal and government731438.04,743N/AAug 6, 2026
text style controloverall731442.263,719N/AAug 6, 2026
text style controlindustry medicine and healthcare731462.64,711N/AAug 6, 2026
text style controlindustry writing and literature and language731421.514,873N/AAug 6, 2026
text style controlexclude ties741441.346,583N/AAug 6, 2026

Pricing

Official vendor API PAYG pricing is summarized first. The table then lists individual provider offerings without treating their minimum as the official price.

Official API
$0.43 input, $2.58 output per 1M
Official provider
Alibaba (China)
Lowest third-party
From $0.3025 input, $1.93 output per 1M via DigitalOcean
Tracked offerings
8
7 rows
Columns

Show columns

DigitalOceanqwen3.5-397b-a17bglobal$0.3025$1.93131.1KAug 7, 2026
CrofAIqwen3.5-397b-a17bglobal$0.35$1.75262.1KAug 7, 2026
Alibaba (China)qwen3.5-397b-a17bglobal$0.43$2.58262.1KAug 7, 2026
Scalewayqwen3.5-397b-a17bglobal$0.60$3.6256KAug 7, 2026
Cortecsqwen3.5-397b-a17bglobal$0.668$4.01262KAug 7, 2026
OVHcloud AI Endpointsqwen3.5-397b-a17bglobal$0.71$4.25262.1KAug 7, 2026
GreenPTqwen3.5-397b-a17bglobal$0.798$4.96262.1KAug 7, 2026

Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.

Qwen3.5 397B A17B versions

All published versions linked to this unique model. The score columns identify the version used by the current overall ranking.

1 rows
Columns

Show columns

Qwen3.5-397B-A17BFeb 16, 202675.2397B262.1K65.5KYesApache 2.0

What is Qwen3.5 397B A17B?

A concise description based on the published model registry.

Qwen3.5-397B-A17B is Qwen's flagship Mixture-of-Experts model with 397 billion total parameters and 17 billion activated parameters. It delivers state-of-the-art performance across knowledge, reasoning, coding, mathematics, multilingual understanding, instruction following, long context, and agent tasks.

Use the benchmark, Arena and pricing sections above as separate evidence. A missing field means the current data snapshot does not support that claim.

Data snapshot: 2026-08-07. Editorial model content is not available in the backend.

Qwen3.5 397B A17B vs nearby models

Open a comparison with the three ranked models immediately above and below this model.

Qwen3.5 397B A17BvsGPT-5.5Qwen3.5 397B A17BvsGPT-5.4Qwen3.5 397B A17BvsDeepSeek-V3.2Qwen3.5 397B A17BvsNova 2 ProQwen3.5 397B A17BvsGemma 4 31BQwen3.5 397B A17BvsClaude Sonnet 3.5

Models similar to Qwen3.5 397B A17B

Recommendations prioritize the same model type and family, then the closest published LLMBoard score.

#23-2.1
AC

Qwen3.7 Plus

Alibaba Cloud / Qwen Team

73.1 LLMBoard

DetailsCompare
#25-2.8
AC

Qwen3.6 Plus

Alibaba Cloud / Qwen Team

72.4 LLMBoard

DetailsCompare
#26-3.0
AC

Qwen3.6 27B

Alibaba Cloud / Qwen Team

72.2 LLMBoard

DetailsCompare
#39-9.5
AC

Qwen3 235B A22B Thinking

Alibaba Cloud / Qwen Team

65.7 LLMBoard

DetailsCompare
#50-13.8
AC

Qwen3.5 122B A10B

Alibaba Cloud / Qwen Team

61.4 LLMBoard

DetailsCompare
#2+15.1
AC

Qwen3.7 Max

Alibaba Cloud / Qwen Team

90.3 LLMBoard

DetailsCompare

FAQ

Common questions about Qwen3.5 397B A17B.

When was Qwen3.5 397B A17B released?

Qwen3.5 397B A17B's default version was released on Feb 16, 2026.

How much does Qwen3.5 397B A17B cost?

Qwen3.5 397B A17B's official API price is $0.43 per million input tokens and $2.58 per million output tokens via Alibaba (China). The lowest tracked third-party offer starts at $0.3025 input and $1.93 output via DigitalOcean.

Who created Qwen3.5 397B A17B?

Qwen3.5 397B A17B is published under Alibaba Cloud / Qwen Team in the model registry.

What is the context window for Qwen3.5 397B A17B?

The default version has a 262.1K token context window.

Is Qwen3.5 397B A17B open weight?

Yes. The default version is marked as open weight under Apache 2.0.

How many API providers offer Qwen3.5 397B A17B?

8 published provider offerings are linked to the default version.

What models should I compare Qwen3.5 397B A17B with?

Nearby ranked alternatives include GPT-5.5, GPT-5.4, DeepSeek-V3.2.

Rankings

OverallCodingText ArenaPricing

Modalities

Image GenerationVideo GenerationSpeech-to-TextEmbeddings

Benchmarks

All BenchmarksReasoningMathCoding

Vendors

All VendorsOpenAIAnthropicGoogle
llmboard.aiCopyright 2026 llmboard.ai