LongCat
By Meituan
Meituan LongCat-2.0: an agent-native coding model with strong code ability
- Pricing:
- Free
Overview
LongCat-2.0 is Meituan's AI Coding Agent model, with trillion-scale total parameters and a 1M long context. It natively supports tool calling and multi-step reasoning to reliably handle long-context agent tasks, excels at code generation, code understanding and automated coding, and integrates deeply with mainstream coding environments such as Claude Code, Hermes, OpenClaw, OpenCode and Kilo Code.
Core capabilities
- Trillion parameters, 1M long context
- native tool calling and multi-step reasoning for long-context agent tasks.
- Superior coding ability
- strong performance on code generation, code understanding and automated programming.
- Deep integration with mainstream IDEs
- works seamlessly with Claude Code, Hermes, OpenClaw, OpenCode and Kilo Code.
Use cases
- Generate functions and complete code during daily coding
- Understand and explain complex logic in unfamiliar codebases
- Run multi-step coding tasks inside automated coding agents
- Analyze large project code and docs with a 1M long context
Who it's for
Software developers, Software engineers, AI coding users, Engineering teams
Pricing
default plan
free
Related tools
VisActorFreeThe smart visualization suite within VisActor that automates chart creation and narrative-driven data storytelling across enterprise and BI scenarios.
Code ArenaFreeAn AI-powered code playground and LLM leaderboard that lets you build full-stack apps, games, and designs while comparing frontier models in real-time.
Visual Studio IntelliCodeFreeAI-driven code completion that lets Visual Studio developers focus on creation
TestDriverPaidAI code review and testing for GitHub — the AI code reviewer that actually runs your app.