Back to News
Meituan•July 6, 2026
LongCat-2.0
Open Source
Large Language Model
Open Source
Agentic Coding
Mixture of Experts
### TL;DR
LongCat-2.0 is a next-generation, trillion-parameter Mixture of Experts (MoE) large language model specifically engineered for agentic coding and complex task execution. It features a native 1-million-token context window and is the first model of its scale to be trained and deployed entirely on a domestic (Chinese) AI ASIC supercomputing cluster.
Key Insights & Metrics
Pricing
Free
Cost structure
Version
2.0
Current release version
Hardware
Recommended 16x H20 GPU setup or equivalent domestic AI ASIC clusters for self-hosted deployment
Compute requirements
Category
Open Source
Licensing model
Region
China
Primary region
Key Features
- 1.6 Trillion parameters with dynamic MoE activation (33B–56B per token)
- Native 1-million-token long context window using LongCat Sparse Attention (LSA)
- Purpose-built for agentic coding and automated task execution
- First trillion-parameter model trained and inferred on domestic AI ASIC superpods
- 135B parameter N-gram Embedding for enhanced architectural efficiency
Discussion
0
Upvotes
0
Downvotes
0 reviews
Sign in to leave a review
Reviews
No reviews yet. Be the first to review!