Skip to main content

Qwen3.6 35B A3B (Reasoning)

AlibabaQwenOpen WeightApache 2.0 · Commercial OK

Description

Qwen3.6-35B-A3B is the first open-weight variant of the Qwen3.6 series, a multimodal Mixture-of-Experts model with 35B total parameters and 3B activated. It pairs a vision encoder with a hybrid 40-layer language model that interleaves Gated DeltaNet linear-attention blocks and Gated Attention blocks (10 × (3 × DeltaNet + 1 × Attention)) over 256 experts (8 routed + 1 shared, expert dim 512). The release prioritizes stability and real-world utility, with substantial gains in agentic coding (frontend workflows, repo-level reasoning) and a new option to preserve reasoning context across turns. Native context length is 262K tokens, extensible to ~1M via YaRN, and the model thinks by default.

Release Date
2026-04-16
Parameters
35.0B
Context Length
262K
Modalities
image, text, video

Capability Radar

70
general
50
coding
80
reasoning
77
scienceest.
50
agents
80
multimodal

Science is estimated from LLM Stats science scores or reasoning when no dedicated science benchmarks are available.

Rankings

Domain#RankScoreSource
Agentic Capability119
38.0
LS
Multimodal Ranking37
60.0
LS

Benchmark Scores (LLM Stats)

(LLM Stats (zeroeval))

Agents

TAU3-Bench67.2%SR
MCP-Mark37.0%SR
VITA-Bench35.6%SR
Toolathlon26.9%SR
DeepPlanning25.9%SR

Code

ZClawBench52.6%SR
Claw-Eval50.0%SR
NL2Repo29.4%SR
SkillsBench28.7%SR

General

C-Eval90.0%SR

Language

MMLU-Redux93.3%SR
MMLU-Pro85.2%SR

Math

AIME 202692.7%SR
HMMT 202590.7%SR
HMMT2589.1%SR
MathVista-Mini86.4%SR
HMMT Feb 2683.6%SR
IMO-AnswerBench78.9%SR

Multimodal

VideoMMMU83.7%SR
VideoMME w/o sub.82.5%SR
MMMU81.7%SR

Reasoning

GPQANYU + Cohere + Anthropic (2023)86.0%SR
LiveCodeBench v680.4%SR
CharXiv-R78.0%SR
SWE-Bench Verified73.4%SR
SWE-bench Multilingual67.2%SR
SuperGPQA64.7%SR
MCP Atlas62.8%SR
Terminal-Bench 2.0Stanford × Laude Institute (2026)51.5%SR
SWE-Bench ProPrinceton NLP (2024)49.5%SR
Humanity's Last Exam21.4%SR

Search

WideSearch60.1%SR

Video

MLVU86.2%SR

Vision

MMBench-V1.192.8%SR
AI2D92.7%SR
RefCOCO-avg0.92 / 100SR
OmniDocBench 1.589.9%SR
VideoMME w sub.86.6%SR
RealWorldQA85.3%SR
EmbSpatialBench0.84 / 100SR
CC-OCR81.9%SR
MMMU-Pro75.3%SR
MVBench74.6%SR
LVBench71.4%SR
Hallusion Bench69.8%SR
RefSpatialBench0.64 / 100SR
SimpleVQA0.59 / 100SR
ODinW50.8%SR
ZEROBench-Sub0.34 / 100SR

AA Evaluation Indices

(Artificial Analysis)

No AA evaluation data available

LLM Stats Category Scores

(LLM Stats (zeroeval))
Language
90
Structured Output
90
Biology
90
Long Context
80
Math
80
Multimodal
80
Physics
80
Spatial Reasoning
80
Embodied
80
Grounding
80
Healthcare
80
Chemistry
80
Text-to-image
80
Video
80
Legal
70
Reasoning
70
Finance
70
Frontend Development
70
General
70
Vision
70
Image To Text
60
Search
60
Economics
60
Code
50
Tool Calling
50
Agents
40

Pricing

Input Price$0.375 / 1M tokens
Output Price$2.25 / 1M tokens
Blended Price (3:1)$0.3125 / 1M tokens

Speed

Tokens/sec118.9
Time to First Token2.02s
Time to Answer47.39s

Provider Price Ranking

Provider Price Ranking

14 providers

Cheapest: engyMost Expensive: Alibaba
ProviderInputOutput
1engyCheapest
$0.045
$0.3
2EmpirioLabs AI
$0.07
$0.42
3SaladCloud AI Gateway
$0.09
$0.6
4DeepInfra
$0.1
$0.95
5Venice AI
$0.1
$1
6ZeroGPU
$0.2
$1.5
7DevPass (LLM Gateway)
$0.248
$1.485
8Opper
$0.248
$1.485
9AIHubMix
$0.254
$1.524
10Cortecs
$0.279
$1.671
11302.AI
$0.28
$1.7
12Scaleway
$0.28038
$1.68225
13GreenPT
$0.342
$2.052
14AlibabaPRIMARY
$0.375
$2.25

Compare pricing across different API providers for this model.

External Sources