LongCat-2.0, a large-scale MoE model with 1.6T total parameters and 48B active parameters

LongCat-2.0 is a large-scale MoE (Mixture-of-Experts) model with a total of 1.6T (1.6 trillion) parameters.

Key features are as follows:

• Active Parameters: Only 48B (48 billion) parameters are activated during inference, enabling efficient computation.
• Architecture: Despite its large-scale parameter count, it adopts an MoE structure to optimize inference cost and speed.
• Scale: Through an expanded parameter scale compared to existing models, it aims for more sophisticated language understanding and generation capabilities.