Skip to main content

Qwen3.6 27B (Reasoning)

AlibabaQwenOpen WeightApache 2.0 · Commercial OK

Description

Qwen3.6-27B is a dense 27-billion-parameter multimodal model in the Qwen3.6 series, supporting both vision-language thinking and non-thinking modes in a single unified checkpoint. The 64-layer language model uses a hybrid layout of 16 repeats of (3 × Gated DeltaNet → FFN, 1 × Gated Attention → FFN) with hidden dim 5120 and FFN intermediate 17408 — Gated DeltaNet has 48/16 heads for V/QK (head dim 128) and Gated Attention has 24/4 heads for Q/KV (head dim 256). It supports a native 262,144-token context extensible to ~1,010,000 via YaRN and is trained with multi-token prediction. The release delivers flagship-level agentic coding, surpassing the previous-generation open-source flagship Qwen3.5-397B-A17B (397B total / 17B active) on every major coding benchmark including SWE-bench Verified (77.2), SWE-bench Pro (53.5), Terminal-Bench 2.0 (59.3), and SkillsBench (48.2), and reaches 87.8 on GPQA Diamond. Released as open weights under Apache 2.0; accessible via Qwen Studio with the Alibaba Cloud Model Studio API coming soon.

Release Date
2026-04-22
Parameters
27.8B
Context Length
262K
Modalities
image, text, video

Capability Radar

80
general
60
coding
80
reasoning
77
scienceest.
60
agents
80
multimodal

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

Rankings

Domain#RankScoreSource
Agentic Capability135
35.0
LS
Multimodal Ranking16
65.0
LS

Benchmark Scores (LLM Stats)

(LLM Stats (zeroeval))

Agents

AndroidWorld70.3%SR

Code

QwenWebBench1487.00 / 2000SR
Claw-Eval60.6%SR
ZClawBench53.4%SR
SkillsBench48.2%SR
NL2Repo36.2%SR

General

C-Eval91.4%SR

Language

MMLU-Redux93.5%SR
MMLU-Pro86.2%SR

Math

AIME 202694.1%SR
HMMT 202593.8%SR
HMMT2590.7%SR
MathVista-Mini87.4%SR
DynaMath85.6%SR
HMMT Feb 2684.3%SR
IMO-AnswerBench80.8%SR

Multimodal

VideoMMMU84.4%SR
MMMU82.9%SR

Reasoning

GPQANYU + Cohere + Anthropic (2023)87.8%SR
LiveCodeBench v683.9%SR
CharXiv-R78.4%SR
SWE-Bench Verified77.2%SR
SWE-bench Multilingual71.3%SR
SuperGPQA66.0%SR
Terminal-Bench 2.0Stanford × Laude Institute (2026)59.3%SR
SWE-Bench ProPrinceton NLP (2024)53.5%SR
Humanity's Last Exam24.0%SR

Video

MLVU86.6%SR

Vision

CountBench0.98 / 100SR
VLMsAreBlind97.0%SR
V*94.7%SR
RefCOCO-avg0.93 / 100SR
MMBench-V1.192.3%SR
OCRBench89.4%SR
VideoMME w sub.87.7%SR
EmbSpatialBench0.85 / 100SR
RealWorldQA84.1%SR
MMStar81.4%SR
CC-OCR81.2%SR
MMMU-Pro75.8%SR
MVBench75.5%SR
RefSpatialBench0.70 / 100SR
ERQA62.5%SR
SimpleVQA0.56 / 100SR

AA Evaluation Indices

(Artificial Analysis)

No AA evaluation data available

LLM Stats Category Scores

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

Pricing

Input Price$0.6 / 1M tokens
Output Price$3.6 / 1M tokens
Blended Price (3:1)$1.04 / 1M tokens

Speed

Tokens/sec55.9
Time to First Token3.52s
Time to Answer104.98s

Provider Price Ranking

Provider Price Ranking

8 providers

Cheapest: DevPass (LLM Gateway)Most Expensive: Novita
ProviderInputOutput
1DevPass (LLM Gateway)Cheapest
$0.12
$1.08
2DeepInfra
$0.32
$3.2
3Venice AI
$0.325
$3.25
4EmpirioLabs AI
$0.41256
$2.47538
5Cortecs
$0.446
$3.008
6OVHcloud AI Endpoints
$0.47
$3.19
7AlibabaPRIMARY
$0.6
$3.6
8Novita
$0.6
$3.6

Compare pricing across different API providers for this model.

External Sources