Qwen (Tongyi Qianwen)

Open-source Large Language Model (LLM) Q LLM Models & Providers

Basic Information

Product Description

Qwen (Tongyi Qianwen) is a series of large language models developed by Alibaba's Tongyi Lab, renowned for its open-source strategy and robust capabilities in both Chinese and English. In early 2026, Qianwen released a major update to the Qwen3.5 series, including Qwen3.5-397B MoE Multi-Chip Edition (the first MoE model released with multi-chip compatibility), Qwen3-Max-Thinking reasoning model, and the AI programming solution Qwen Coding Plan. Qianwen leads the Chinese enterprise-level large model invocation market with a 32% market share.

Core Features/Characteristics

  • Qwen3.5-397B MoE: Full parameter count of 397 billion, activated parameters of 17 billion, the first MoE model released with multi-chip compatibility
  • Qwen3.5-Plus: Supports text, image, and video input, performance comparable to Qwen3 Max
  • Qwen3-Max-Thinking: Flagship reasoning model combining thinking and non-thinking modes
  • Qwen3-VL-Flash: Visual model suitable for scenarios like security, inspection, and photo problem-solving
  • Qwen3.5-Flash: Fast reasoning model with flexible tiered pricing
  • Qwen Coding Plan: AI programming solution
  • Multimodal Capabilities: Covers text, images, videos, and audio
  • Open-source Ecosystem: Core models are open-source, freely available to the community

Business Model

  • Open-source and Free: Multiple model versions are open-source and can be deployed locally
  • Alibaba Cloud Bailian Platform: API services with tiered pricing based on token usage
  • Enterprise Customization: Enterprise-level solutions based on Tongyi Qianwen
  • Tongyi Qianwen App: Free AI assistant for consumers

Target Users

  • Enterprise users in the Chinese market (leading with 32% market share)
  • Open-source community developers
  • Developers and enterprises within the Alibaba Cloud ecosystem
  • AI application scenarios requiring Chinese language optimization
  • Visual AI applications (security, inspection, etc.)

Competitive Advantages

  • Leading market share in Chinese enterprise-level large models (32%)
  • Strong infrastructure support from Alibaba Cloud
  • Open-source strategy wins developer community
  • Excellent Chinese language understanding
  • Complete model matrix, covering from lightweight to flagship models
  • MoE architecture enables efficient reasoning
  • Multi-chip compatibility supports deployment on domestic chips

Market Performance

  • Leads the Chinese enterprise-level large model invocation market with a 32% share
  • One of the most downloaded Chinese open-source LLMs on Hugging Face
  • Qwen3.5 ranks first in multiple global benchmark tests
  • Continuous growth of enterprise customers on Alibaba Cloud Bailian Platform
  • High activity in the open-source community, with numerous derived fine-tuned versions

Relationship with the OpenClaw Ecosystem

Qwen is the most important Chinese open-source LLM in the OpenClaw ecosystem. Users can run Qwen models locally using tools like Ollama (with various parameter sizes available) or access them via Alibaba Cloud Bailian API. Qwen's open-source strategy and rich model matrix provide OpenClaw users with flexible choices—from the lightweight Qwen3.5-Flash to the flagship Qwen3.5-397B, catering to different hardware conditions and usage scenarios.

External References

Learn more from these authoritative sources: