NEWHow AI models draw the world map · Gecko Tests
Beta
Home/Models/Qwen3.5 397B A17B
Alibaba Qwen logo

Qwen3.5 397B A17B

by Alibaba Qwen · Released Feb 2026

Open SourceMultimodal
69.3
avg score
Rank #44
Compare
Better than 87% of all models
Context
262K tokens (~131 books)
Input $/1M
$0.60
Output $/1M
$3.60
Type
multimodal
License
Open Source
Benchmarks
29 tested
Data as of
About

The Qwen3.5 series 397B-A17B native vision-language model is built on a hybrid architecture that integrates a linear attention mechanism with a sparse mixture-of-experts model, achieving higher inference efficiency. It delivers...

Tested on 29 benchmarks · BenchGecko score 69.3. Top scores: Chatbot Arena Elo · Overall (1441.8%), Chatbot Arena Elo · Coding (1399.4%), OpenCompass · AIME2025 (92.3%).

Capabilities
coding
83.0
#4 globally
math
71.3
#49 globally
knowledge
59.4
#64 globally
agentic
24.9
#56 globally
speed
41.8
#62 globally
general
44.5
#60 globally
language
91.5
#10 globally
Benchmark Scores
Compare All
Tested on 29 benchmarks · Ranked across 8 categories
Score Distribution (all 326 models)
0255075100
▲ You are here
codingCompare coding →
OpenCompass · LiveCodeBenchV6

OpenCompass Live Code Bench v6. Fresh competitive programming problems to evaluate code generation without memorization.

83.0·
mathCompare math →
OpenCompass · AIME2025

OpenCompass evaluation on AIME 2025 problems. Tests mathematical reasoning on fresh competition problems.

92.3·
OTIS Mock AIME 2024-2025

Mock AIME (American Invitational Mathematics Exam) problems from OTIS. Tests mathematical competition performance.

88.9·
FrontierMath-Tiers-1-3-v2-Private
32.9·
knowledgeCompare knowledge →
OpenCompass · GPQA-Diamond

OpenCompass evaluation of GPQA Diamond. PhD-level science questions from the hardest subset.

88.4·
OpenCompass · MMLU-Pro

OpenCompass MMLU-Pro evaluation. Harder knowledge test with more answer choices.

87.6·
GPQA diamond

Graduate-level science questions written by PhD experts. Diamond subset contains questions where experts disagree, testing deep understanding.

85.2·
Quick compare:OpenAI logovs GPT-5.3-CodexAnthropic logovs Claude Opus 4.8Microsoft logovs phi-3-medium 14B
Excellent (85+) Good (70-85) Average (50-70) Below (<50)
Model Family · Alibaba Qwen Qwen 3.5
Feb 2026
69.3
$0.60/M in262Kctx29 benchmarks
Feb 2026
$0.26/M in(-0.34)1.0Mctx(+738K)
Apr 2026
$0.30/M in(+0.04)1.0Mctx
Feb 2026
$0.26/M in(-0.04)262Kctx(-738K)15 benchmarks
Feb 2026
$0.26/M in262Kctx5 benchmarks
Feb 2026
53.2-16.1
$0.14/M in(-0.12)262Kctx10 benchmarks
Mar 2026
39.7-13.5
$0.17/M in(+0.03)262Kctx19 benchmarks
Mar 2026
$0.17/M in262Kctx
Feb 2026
34.7-5.0
$0.07/M in(-0.11)1.0Mctx(+738K)12 benchmarks
See the full Qwen 3.5 family →
Recently Happened
Qwen3.5 397B A17B pricing increased 20%
Nov 9, 2026
Qwen3.5 397B A17B pricing increased 54%
Oct 22, 2026
Qwen3.5 397B A17B pricing dropped 35%
Oct 20, 2026
Qwen3.5 397B A17B pricing dropped 9%
Oct 17, 2026
Qwen3.5 397B A17B pricing increased 50%
Sep 9, 2026
Qwen3.5 397B A17B pricing dropped 33%
Sep 7, 2026
Similar Models
OpenAI logo
GPT-5.3-Codex
OpenAI
69.3$1.75/1M
Anthropic logo
Claude Opus 4.8
Anthropic
69.2$5.00/1M
Microsoft logo
phi-3-medium 14B
Microsoft
69.2n/a
Links
Info
Alibaba QwenPricing explorerDevelopers · API
Research
Technical Report
Documentation
API DocsPlayground
Community
@Alibaba Qwen
Source Code
GitHubHugging Face
BenchGecko API
qwen3-5-397b-a17b
Specifications
  • Typemultimodal
  • Context262K tokens (~131 books)
  • ReleasedFeb 2026
  • LicenseOpen Source
  • StatusActive
  • Cost / Message~$0.005
Available On
Alibaba Qwen logoAlibaba Qwen$0.60
Categories
codingmathknowledgeagenticspeedgenerallanguage
Learn More
context-windowtransformeropen weightstokens
Share & Export
Tweet
Related Models
OpenAI logoGPT-5.3-Codex
69.3
Anthropic logoClaude Opus 4.8
69.2
Microsoft logophi-3-medium 14B
69.2
OpenAI logoGPT-5.2 Pro
69.8
Qwen3.5 397B A17B is an open-source multimodal AI model by Alibaba Qwen, released in February 2026. It has an average benchmark score of 69.3. Context window: 262K tokens.

Related Models

GPT-5.3-Codex · OpenAIClaude Opus 4.8 · Anthropicphi-3-medium 14B · MicrosoftGPT-5.2 Pro · OpenAIClaude Instant · Anthropic

Benchmarks

Chatbot Arena Elo · OverallChatbot Arena Elo · CodingOpenCompass · AIME2025OpenCompass · IFEvalOTIS Mock AIME 2024-2025

Related Pages

Alibaba Qwen · ProviderAlibaba Qwen · EconomyAll ModelsCompare ModelsPricingDevelopers · APIContext Window · Glossary

Key facts · as of 2026-11-11

  • Qwen3.5 397B A17B by Alibaba Qwen. BenchGecko score 69.3, rank 43 of 326 scored models (normalized average of public benchmark scores).
  • List price $0.60 input · $3.60 output per 1M tokens (as of 2026-11-11).
  • Sold by 8 providers (as of 2026-11-11): Alibaba $0.39 in / $2.34 out · DeepInfra (fp8) $0.45 in / $3.00 out · Parasail (fp8) $0.50 in / $3.60 out · AtlasCloud (fp8) $0.55 in / $3.50 out · GMICloud (fp8) $0.60 in / $3.60 out · and 3 more. Every provider

How to cite · data as of 2026-11-11

Qwen3.5 397B A17B · benchmarks, pricing and providers. BenchGecko, data as of 2026-11-11. https://benchgecko.ai/model/qwen3-5-397b-a17b

Credit "Source: BenchGecko" with a link. Prices per provider and Gecko Tests are BenchGecko data (CC BY 4.0); benchmark scores keep their original source, listed in the JSON. JSON · llms.txt · MCP