Skip to main content
9 – 17 UHR +49 8031 3508270 LUITPOLDSTR. 9, 83022 ROSENHEIM
DE / EN
LLM Alibaba China

Alibaba Qwen

Alibaba Qwen 3.6 - Agentic real-world AI with MoE efficiency. Qwen3.6-35B-A3B as new top pick (Apache 2.0). AI consulting from Germany for GDPR-compliant self-hosting.

License Apache 2.0
GDPR Hosting Available
Context 128k (standard), up to 1M (Qwen3.5-397B-A17B) Tokens
Modality Text, Image, Video → Text, Image

Versions

Overview of available model variants

ModelReleaseEUStrengthsWeaknessesStatus
Qwen3.6-Max-Preview
2026-04
Most powerful Qwen model (proprietary) Multimodal (Text + Image) Agentic real-world workflows
Preview status API-only, not open source
Current
Qwen3.6-Plus
2026-04
Focus on real-world agents Multimodal
Proprietary, not open source
Current
Qwen3.6-27B
2026-04
28B dense model Open weights Multimodal (Text + Image)
High resource needs for 27B class
Current
Qwen3.6-35B-A3B Recommended
2026-04
MoE: 36B total, only 3B active – extremely efficient Apache 2.0 license Multimodal Outperforms Google Gemma 4 in benchmarks
Current
Qwen3.5-397B-A17B
2026-02
Flagship (MoE, 17B active) Native multimodality (Text, Image, Video) Up to 1M context Agentic AI workflows #5 globally in math reasoning (LM Arena)
High resource needs
Current
Qwen3.5-Max-Preview
2026-03
Leading Chinese model on LM Arena #5 globally in math reasoning
Preview status
Current
Qwen3.5-9B
2026-03
Strong logical reasoning Rivals 30B+ models Natively multimodal Ideal for edge servers
Limited on very complex tasks
Current
Qwen3.5-4B
2026-03
Native multimodal agent UI navigation and document analysis Compact and efficient
Limited capacity
Current
Qwen3.5-2B
2026-03
Optimized for mobile and edge Natively multimodal
Limited capacity
Current
Qwen3.5-0.8B
2026-03
Ultra-compact for IoT and smartphones Natively multimodal
Strongly limited capacity
Current
Qwen3-235B-A22B
2025
Strong predecessor (MoE, 22B active) Open weights Up to 1M context (configuration)
High resource needs
Current
Qwen3-32B
2025
General-purpose Open weights
Current

Use Cases

Typical applications for this model

Multilingual Communication
Video Analysis and Summarization
Image Generation with Text
Self-Hosted AI
Agentic Workflows & Automation
International Teams (Asia-Europe)
Content Creation
Safety & Moderation
Translation

Technical Details

API, features and capabilities

API & Availability
Availability Public
Latency (TTFT) Varies (Self-Hosted)
Features & Capabilities
Tool Use Function Calling Structured Output Vision Reasoning Mode File Upload
Training & Knowledge
Knowledge Cutoff 2025-12
Fine-Tuning Available (LoRA, Full, PEFT)
Language Support
Best Quality Chinese, English, German
Supported 200+ languages and dialects (Qwen3.5)
Outstanding in Chinese, very good in European languages, expanded coverage for South Asia, Africa, and Oceania

Hosting & Compliance

GDPR-compliant hosting options and licensing

GDPR-Compliant Hosting Options
Self-Hosted
Own Infrastructure
Recommended for enterprise data
AWS Bedrock
Frankfurt (eu-central-1)
Qwen3 models available
Alibaba Cloud
Frankfurt (eu-central-1)
Alibaba Cloud EU - verify data protection
License & Hosting
License Apache 2.0
Security Filters Customizable (Qwen3Guard)
On-Premise Edge-capable

Benchmarks

Performance comparison with standardized tests

MMLU
90.6%
HumanEval
93%
GSM8K
79.3%
C-Eval
85.6%

As AI consultants based in Rosenheim, Germany, we recommend Qwen models for enterprises looking for an open-source alternative with Apache 2.0 license. With the Qwen 3.6 generation (April 2026), Alibaba focuses on agentic real-world workflows – from efficient MoE models to proprietary flagships. For GDPR-compliant usage in the DACH region (Germany, Austria, Switzerland), we recommend self-hosting or AWS Bedrock in Frankfurt.

Qwen 3.6: Agentic Real-World AI (April 2026)

In April 2026, Alibaba released the Qwen 3.6 generation – focused on real-world agent workflows:

Qwen3.6-35B-A3B – Our New Top Pick

The most efficient model in the family uses MoE architecture (36B total, only 3B active) and outperforms Google Gemma 4 in benchmarks. The Apache 2.0 license enables unrestricted self-hosting.

Qwen3.6-27B – Dense Alternative

With 28 billion parameters, this dense model delivers strong performance for applications that don’t support MoE architecture.

Qwen3.6-Plus and Max-Preview

The proprietary models are only available via API and reflect Alibaba’s strategic shift toward commercial offerings. Qwen3.6-Plus focuses on real-world agents – autonomous AI that executes real tasks like app control and document editing.

Note: With Qwen 3.6, Alibaba introduces proprietary models for the first time that are not available as open source. For self-hosting, we recommend Qwen3.6-35B-A3B (Apache 2.0).

Qwen 3.5: The Next Generation

Native Multimodality

Qwen 3.5 unifies text, image, and video in one architecture:

  • Video Analysis: Understands up to 2 hours of video in a single prompt
  • Timestamp-Precise: Identifies events at second-level resolution
  • Long Context: Up to 1 million tokens (entire books, large codebases)
  • Flexible Input: URLs, local files, frame sequences

Agentic AI & Automation

Qwen 3.5 can execute autonomous workflows:

  • App interaction on smartphones
  • Document editing and email management
  • Travel booking and process automation
  • Multi-step tasks with tool use

Efficiency through Mixture-of-Experts

The Qwen3.5-397B-A17B model uses:

  • 397 billion parameters total
  • Only 17 billion active per inference
  • 60% lower costs than predecessors
  • 8-19x higher throughput than Qwen3

New: Qwen 3.5 Small Model Series (March 2026)

In March 2026, Alibaba released a new series of compact models for edge and mobile applications:

ModelParametersUse Case
Qwen3.5-9B9BEdge servers, rivals 30B+ models
Qwen3.5-4B4BUI navigation, document analysis
Qwen3.5-2B2BMobile devices
Qwen3.5-0.8B0.8BIoT, smartphones

All small models are natively multimodal and agent-capable. They are particularly suited for on-device AI where privacy is ensured through local processing.

Qwen3.5-Max-Preview

With Qwen3.5-Max-Preview, Alibaba leads the Chinese AI rankings on LM Arena and achieves 5th place globally in math reasoning.

Key Strengths

Open Source & Apache 2.0

  • Full Control: Model runs in your infrastructure
  • No License Costs: Commercial use permitted
  • Customizable: Fine-tuning on your own data possible
  • GDPR-Friendly: No data leaves your company

Expanded Multilingual Support

Qwen 3.5 supports 200+ languages and dialects:

  • Chinese (outstanding)
  • European languages (very good)
  • Expanded coverage: South Asia, Africa, Oceania
  • Competitive with Western models

Text-in-Image Generation

Qwen-Image is leading in:

  • Complex text layouts
  • Multilingual text rendering
  • Paragraph-level semantics
  • Fine detail work

Availability

AWS Bedrock (EU)

Qwen3 models are now available on AWS Bedrock in Frankfurt:

  • Fully managed and serverless
  • EU data residency (GDPR-compliant)
  • Integration with AWS services
  • Qwen3-32B, Qwen3-235B, Qwen3-Coder available

Self-Hosting

All Qwen models can be operated in your own infrastructure - this way all data remains under your control.

Benchmarks & Performance

BenchmarkScoreModel
MMLU90.6%Qwen3-235B VL
HumanEval93%Qwen3-32B
GSM8K79.3%Qwen3-32B
C-Eval (CN)85.6%Qwen3-32B

Qwen 3.5 competes with GPT-4-class models and clearly outperforms other open-source alternatives.

Hardware Requirements (Self-Hosted)

ModelVRAMRecommended GPU
Qwen3.5-397B-A17B80+ GBH100/MI300X
Qwen3-235B-A22B48+ GBA100/H100
Qwen-Image 20B48+ GBA100/H100
Qwen3 (smaller variants)16-24 GBRTX 4090

Integration with CompanyGPT

Qwen models can be integrated into CompanyGPT as a self-hosted option or via AWS Bedrock - full GDPR compliance guaranteed.

Our Recommendation

Qwen 3.6 sets new standards for agentic AI with real-world workflows. For DACH enterprises, we recommend:

  • Qwen3.6-35B-A3B: New top pick – extremely efficient (3B active), Apache 2.0, outperforms Gemma 4
  • AWS Bedrock Frankfurt: For managed solution with EU data residency
  • Self-Hosting: For maximum data control and customizability
  • Qwen3.5-397B-A17B: For highest multimodality requirements

Consultation for this model?

We help you select and integrate the right AI model for your use case.