Alibaba Qwen

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 3.7 generation (May/June 2026), Alibaba ships a new flagship with 1M context and full agent workflows – while Qwen 3.6 continues to serve as the open-weight line. For GDPR-compliant usage in the DACH region (Germany, Austria, Switzerland), we recommend self-hosting or AWS Bedrock in Frankfurt.

Qwen 3.7: The New Flagship (May/June 2026)

At the Apsara Summit in Hangzhou on 20 May 2026, Alibaba unveiled Qwen3.7-Max – the new proprietary top model of the Qwen family. On 1 June 2026, Qwen3.7-Plus followed as the multimodal agent variant at roughly one-sixth of the per-token cost.

Qwen3.7-Max

• 1 million token context window – fits large-scale agent workflows with long documents
• Top scores (per Alibaba) on Terminal-Bench 2.0, SWE-Bench Pro and MCP-Atlas
• Support for the Anthropic protocol (MCP)
• Available via Alibaba Cloud Model Studio (proprietary, no open weights)

Qwen3.7-Plus

• Multimodal agent (text, image, video) on the Qwen 3.7 backbone
• Deep reasoning, tool invocation and autonomous iteration via the Bailian platform
• Listed at roughly 1/6 of Qwen3.7-Max pricing – attractive for vision-heavy agent workloads
• First seen as preview on LM Arena (around 14 May 2026), GA on 1 June 2026

Note: As of June 2026, Qwen 3.7 is API-only via Alibaba Cloud. For self-hosting, Qwen3.6-35B-A3B (Apache 2.0) remains our recommendation.

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 Agent & agentic Qwen workflows

Beyond the models themselves, the Qwen team maintains Qwen-Agent (github.com/QwenLM/Qwen-Agent), an open-source framework for building AI agents on top of the Qwen models (Qwen 3.0 and newer). It encapsulates the typical building blocks of production agents and removes most of the boilerplate for developers:

• Function calling / tool use: native support for parallel, multi-step, and multi-turn tool calls, including automatic parsing and execution
• MCP integration: connect external tools and data sources (e.g. GitHub, filesystem, SQLite) via the open Model Context Protocol (MCP) standard
• Code interpreter: execute Python code in an isolated Docker sandbox
• RAG over long contexts: document QA over 1M+ tokens using hybrid RAG and agent-based decomposition
• BrowserQwen: a browser assistant with a Chrome extension as a reference implementation

The framework supports Qwen3, Qwen3-VL, Qwen3-Omni, Qwen3-Coder and QwQ, among others, and can be run either via the Alibaba Cloud API or against local, OpenAI-compatible servers (vLLM, SGLang).

On the model side, Alibaba is pushing the agentic direction hard. Qwen3.7-Max is explicitly positioned as the “Agent Frontier”: Alibaba demonstrates (as of June 2026) long autonomous runs – for instance a 35-hour kernel-optimization run with over 1,000 tool calls – and reports top scores on tool-use benchmarks such as MCP-Mark and Terminal-Bench 2.0. According to the vendor, the model serves as a drop-in backbone for agent frameworks (including Claude Code, Qwen Code, OpenClaw) and supports MCP. On the open-weight side, Qwen3.6-35B-A3B targets agentic coding explicitly and works, for example, with the open-source terminal agent OpenClaw.

Sovereignty & GDPR: The cleanest path for agentic Qwen workflows in the DACH region is the combination of open Qwen weights (e.g. Qwen3.6-35B-A3B, Apache 2.0) plus Qwen-Agent, fully self-hosted on EU infrastructure – so neither prompts nor tool data leave your company. Managed options are Qwen3 models on AWS Bedrock (Frankfurt) or Azure AI Foundry. The newest closed flagships (Qwen 3.7-Max/Plus), by contrast, are (as of June 2026) API-only via Alibaba Cloud – here the data-protection situation must be assessed separately.

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:

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

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

Hardware Requirements (Self-Hosted)

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

With Qwen 3.7, Alibaba ships a new proprietary flagship with 1M context and a full agent stack. On the open-source side, Qwen 3.6 remains the benchmark. For DACH enterprises, we recommend:

• Qwen3.6-35B-A3B: Top pick for self-hosting – extremely efficient (3B active), Apache 2.0, outperforms Gemma 4
• Qwen3.7-Max / Plus: For maximum agent performance and 1M context – only via Alibaba Cloud Model Studio
• AWS Bedrock Frankfurt: For managed Qwen3 models with EU data residency
• Self-Hosting: For maximum data control and customizability
• Qwen3.5-397B-A17B: For the highest open-source multimodality requirements