1M
Max Context
0+
Available Models
64K
Max Output
Starting at ¥0.2
Per Million Tokens
Model Families
From flagships to lightweight, from general to specialized — meeting every scenario's needs.
Qwen3.7 Series
Latest FlagshipThe top-tier flagship released in 2026, comprehensively upgraded for the AI agent era. Coding, office tasks, and long-cycle autonomous execution capabilities significantly improved. Supports thinking mode switching, function calling, and web search. Million-level context window.
Qwen3.7 Max
Qwen3.6 Series
Flagship RecommendedFlagship series released in 2026, achieving comprehensive leaps in reasoning depth, coding capability, and multimodal understanding. Max Preview is the current Qwen flagship, while Plus balances performance and cost.
Qwen3.6 Max PreviewQwen3.6 Plus
Qwen3.5 Series
Production PowerhouseProduction-grade series validated through large-scale online testing. Plus is the balanced mainstay with million-level context, while Flash delivers high-speed responses at extremely low cost, ideal for high-concurrency production scenarios.
Qwen3.5 PlusQwen3.5 Flash
Qwen3 Coding Series
Code SpecializedCoding models optimized for software development, reaching top-tier performance in code completion, refactoring, debugging, and multi-file understanding. Coder Plus is suited for complex engineering tasks, while Coder Flash is ideal for real-time coding assistance.
Qwen3 Coder PlusQwen3 Coder Flash
Multimodal Series
Visual UnderstandingMultimodal models that understand both text and images simultaneously. Supports image analysis, OCR, chart interpretation, visual Q&A, and more. VL Plus is suited for high-precision tasks, while VL Flash handles real-time image processing.
Qwen3 VL PlusQwen3 VL FlashQwen3 Omni Flash
Core Capabilities
Key technical advantages of Qwen series models.
Million-Level Context
Qwen3.5/3.6 series supports up to 1 million Token ultra-long context windows, capable of processing entire books, complete code repositories, or hours of conversation history in a single pass without segmentation.
Superior Coding Capability
From simple scripts to complex system design, Qwen consistently leads in code generation, completion, refactoring, and debugging. The specialized Coder series ranks among the top in HumanEval, MBPP, and other coding benchmarks.
Function Calling & Tool Use
Native support for Function Calling and tool invocation protocols. Seamlessly integrate with external APIs, databases, search engines, and other services to easily build complex AI Agent workflows.
Multimodal Understanding
VL series models handle both text and image inputs simultaneously, precisely completing image description, chart interpretation, document OCR, visual reasoning, and more — bridging the boundary between vision and language.
Technical Highlights
Qwen's differentiated advantages in core technology.
Top-Tier Chinese Capability
As Alibaba's self-developed large language model, Qwen has a natural advantage in Chinese understanding, generation, and dialogue. Whether it's long-form writing, professional translation, or cultural context comprehension, it consistently delivers outstanding performance.
Thinking Mode
Qwen3.6 series introduces deep thinking mode. Before answering complex questions, the model first performs internal reasoning chain analysis, significantly improving accuracy in mathematical proofs, logical reasoning, and multi-step planning.
Extreme Cost-Efficiency
From Flash series starting as low as ¥0.2/million Tokens, to the top-tier performance of the flagship Max series, Qwen offers a complete pricing gradient from budget to premium, satisfying diverse budget requirements.
Multi-Protocol Compatible Integration
Through NexusFlow's unified integration, Qwen text models support three public protocols: OpenAI Chat, Anthropic Messages, and Gemini-compatible. Existing OpenAI, Anthropic, or Google GenAI clients can all migrate using their respective protocols.
Use Cases
Qwen is driving the intelligent transformation of these fields.
Intelligent Customer Service & Chat
Million-level context + ultra-fast response. The Flash series is the ideal choice for high-concurrency online customer service chatbots.
Code Development Assistant
The Coder series deeply understands code logic, from code completion to architecture design, comprehensively boosting development efficiency.
Long Document Analysis
The million-Token context window can process entire contracts, research reports, or technical documents in a single pass without segmentation.
Multimodal Applications
VL series supports mixed text-image input, suited for e-commerce image understanding, document OCR, medical imaging assistance, and more.
AI Agent Development
Native function calling + tool use capabilities make it easy to build autonomous agents that call external services.
Content Creation
Outstanding Chinese writing capabilities, covering marketing copy, technical documentation, creative writing, and all kinds of content generation needs.