Compare AIFind AIAI NewsAI How-To
About Us
PrivacyTermsFAQContactContact
AIB Inc.Company info
© 2026 AIB Inc.
Alibaba
Alibaba

Qwen3.6 35B A3B

Release Date 2026-04-27(Knowledge cutoff 2026-04-01)··262K IN · 66K OUTProprietary Model|API available
Overall2.5Top 64%Automation2.0Top 81%Office2.5Top 61%Coding2.0Top 80%

Qwen3.6-35B-A3B is an open-weight multimodal model from Alibaba Cloud with 35 billion total parameters and 3 billion active parameters per token. It uses a hybrid sparse mixture-of-experts architecture combining Gated Linear Units and advanced attention mechanisms to deliver high-performance reasoning and multimodal understanding.

  1. 01ImagesGenerated images
  2. 02CostCost · Usage · Speed
  3. 03User CompareUser Comparisons
  4. 04Expert EvalExpert Performance Evaluation
  5. 05OpinionPublic Opinion
  6. 06Related Info

API cost

Market price (median)
$0.125IN$1OUT
Price range10 providers
$0.05~$0.25IN$0.7~$1.8OUT
Calculate Cost
IN
IN
$0.125-25.6%
$0.125-25.6%
OUT
OUT
$1-5.4%
$1-5.4%
9/2310/8
Calculate Cost
SourceOpenRouterLast updatedOct 7, 2026

Output Speed

Fastper 2,000 tokens
First Output0.6s
total25.0s
82tok/s
First Output 0.6stotal: 25.0s
SourceOpenRouterLast updatedOct 8, 2026

Expert Performance Evaluation

Overall performance

MoreAll overall performance AI rankings
QHow strong is it across many tests combined?AA Intelligence Index·Bottom 25%#89
18.2Reasoning effort: Thinking

By effortNone 15.2Thinking 18.2

About AA Intelligence IndexMore about AA Intelligence Index

Automation (agents)

MoreAll automation AI rankings
9.3%Reasoning effort: Thinking95.3%62.8%

By effortNone 5.4%Thinking 9.3%

About τ³-BankingMore about τ³-Banking

Documents and office work

MoreAll documents and office work AI rankings
QCan it pull scattered facts from several long documents into one answer?Long Context Reasoning·Bottom 32%#81
71.7%Reasoning effort: Thinking

By effortNone 64.3%Thinking 71.7%

About Long Context ReasoningMore about Long Context Reasoning

Coding

MoreAll coding AI rankings
0.0%Reasoning effort: Thinking49.5%44.9%34.8%41.936.6%34.5%23.0%

About Terminal-Bench 4.0More about Terminal-Bench 4.0

Writing and chat

MoreAll writing and chat AI rankings
QDoes it follow tricky rules like format and length without missing any?IFBench·Bottom 42%#48
64.4%Reasoning effort: Thinking

By effortNone 36.2%Thinking 64.4%

About IFBenchMore about IFBench

Math and science

MoreAll math and science AI rankings
22.2%Reasoning effort: Thinking0.3%84.1%86.7%20.4%78.9%44.4%26.0%22.0%73.9%29.7

By effortNone 13.9%Thinking 22.2%

About Humanity's Last ExamMore about Humanity's Last Exam

Image understanding

MoreAll image understanding and analysis AI rankings
78.0%Self-reported by developer75.3%

About CharXivMore about CharXiv

Specialized fields

95.1%Reasoning effort: Thinking80.8%83.5%79.8%90.2%74.6%

About Japanese MT-BenchMore about Japanese MT-Bench

See each evaluation’s detail page for its source and last update·Data sources & removal requests

Related Info

Alibaba
Public
Alibaba
🇨🇳 Hangzhou2023$267.3B~220,000 (total)
Qwen3.8 FlashQwen Image 3 ProQwen3.7 PlusQwen3.7 MaxQwen3.5 Plus