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

Announced on Hacker News: LongCat-2.0, a new AI model with a mixture-of-experts architecture. The model features one point six trillion total parameters, but only forty-eight billion are active during any given query. This design improves computational efficiency compared to traditional dense models at similar scales. It represents ongoing progress in building capable language models that require lower computational overhead.

Listen to this story

Hear this and more stories in a personalized audio briefing.

Open The Chonkerton