Alibaba

Qoder CLI is an AI-powered coding assistant for intelligent code generation, debugging, and development workflows. Features TUI mode, inline chat, and project management.

Qwen Code is Alibaba's open source terminal coding agent for understanding large codebases, editing files, executing commands, and connecting to Qwen or OpenAI-compatible models.

Qoder Desktop is an autonomous development desktop with coding agents, deep codebase awareness, intelligent autocomplete, and Editor and Quest workspaces.

An agentic coding platform that brings autonomous AI agents to your IDE. Features code completion, Ask mode for debugging, and Agent mode for complex multi-step development tasks.

Qwen3.6-27B

Dense open-weight Qwen3.6 model for multimodal reasoning and agentic coding, with 27.8B parameters, hybrid thinking, and a 262K-token context.

Qwen3.6-35B-A3B

Open-weight Qwen3.6 MoE model for agentic coding and multimodal reasoning, with 35B total and 3B active parameters, offered through the API as Qwen3.6-Flash.

Qwen3.6-Max-Preview

Proprietary Qwen3.6 flagship preview for agentic coding, knowledge, and instruction following, later superseded by Qwen3.7-Max.

Qwen3.6 Plus

Multimodal Qwen model offered in Alibaba Cloud Coding Plan for coding agents and long-context development tasks.

Qwen3.7-Max

Alibaba's Qwen3.7 flagship for programming, office productivity, tool use, and long-horizon autonomous tasks with a 1M-token context window.

Qwen3.7 Plus

Alibaba's multimodal Qwen model balancing coding, reasoning, speed, and cost with a 1M-token context window.

Qwen3 Coder 30B-A3B

Qwen3 Coder is a specialized code generation model with 32B parameters and 32K context window, optimized for efficient code writing and debugging tasks.

Qwen3 Coder 480B-A35B

Qwen3 Coder is a specialized code generation model with 32B parameters and 32K context window, designed for software development and programming assistance.

Qwen3-Coder-Next

Alibaba's sparse open-weight coding model with 80B total and 3B active parameters, built for efficient long-context coding agents.