LongCat-2.0
Open Source
1.6T MoE trained entirely on AI ASICs
LongCat-2.0 is an MIT-licensed 1.6T-parameter MoE model with ~48B active parameters, 1M context, LongCat Sparse Attention, and post-training for coding and agentic workflows. It was trained on AI ASIC superpods and integrates with Claude Code, OpenClaw, and Hermes.
LongCat-2.0 is a Open Source company. It launched on Product Hunt on Jul 7, 2026, with 148 upvotes.
LongCat-2.0 timeline
- 1.6T MoE trained entirely on AI ASICs
People at LongCat-2.0
No one on file at LongCat-2.0 yet. New people are added as we find them.