Qwen3.7 Max

Coding Agent：Sets new global records across multiple benchmarks in programming evaluations, leading in China.

Deep Reasoning：Delivers outstanding performance in mathematical reasoning and similar tests.

Long-Horizon Autonomous Execution：Supports 35 hours of continuous operation and over a thousand tool calls, capable of independently completing complex projects.

1M Ultra-Long Context：Supports a 1-million-token context window, enabling processing of large codebases or lengthy documents.

Users pursuing ultimate text reasoning：A pure-text flagship with approximately 1.2T parameters, focused on high-intensity agent tasks, million-line code refactoring, and high-precision financial/legal reasoning scenarios.

Core business system developers：For scenarios where stability and precision are paramount (e.g., financial risk control, legal documents, research papers), Max is the preferred choice.

AI application developers：Both models natively support Function Calling and MCP protocols, with compatibility for mainstream frameworks like Claude Code, enabling rapid agent development.

Enterprise digital transformation users：Covers diverse scenarios including automated office work, e-commerce marketing, and data analytics, with private deployment and fine-tuning capabilities available.