Meituan's Longcat Launches LongCat-2.5 Native Multimodal AI Model: 1.6T Parameters for Long-Horizon Agent Tasks
Meituan's Longcat released LongCat-2.5, a native multimodal model with 1.6 trillion total parameters and about 48 billion activated per token, built for long-horizon agent tasks. It strengthens native visual understanding and agent programming through more long-trajectory training data and large-scale long-horizon reinforcement learning, representing visual signals as multi-level discrete tokens. The release matters because it targets end-to-end engineering work such as developing complex codebases and building complete projects, not just short chat responses.
SOURCES
- IT之家2026-11-02 01:18