>
L
Model Details
Domain:
Model Access:
Open weights (unrestricted)
AI Tools for You
Introduction
We are introducing and open sourcing LongCat-2.0, a large-scale MoE language model with 1.6 trillion total parameters and ~48 billion activated per token — a substantial step up from previous LongCat models, accompanied by several architectural improvements.
Benchmarking
FLOPs1.008e+25
Notes: 6 * 48e9 active parameters * 35e12 tokens = 1.008e25 FLOP. Confidence between Likely and Speculative because the 35T is a lower bound and 48B is an aggregate metric since expert size isn't consistently 48B. Hardware quantity of 50k is also a lower bound - "LongCat-2.0 is pre-trained on over 50K AI ASICs, "
Training
Training Code AccessibilityMIT license
Hardware Quantity50000
Parameters
Parameters1600000000000
Notes: 1.6T total, 48B active
Related ModelsView all models
