LongCat-2.0: A 1.6-Trillion-Parameter Model Trained on Home-Grown Silicon
LongCat-2.0, a 1.6-trillion-parameter model, is the first of its size to be trained end-to-end on home-grown silicon, according to the company, serving as a direct response to US export controls. The most remarkable aspect of Meituan’s new artificial intelligence model is not merely its immense scale, though it is indeed substantial, but rather the hardware it was developed on.
The Chinese […]
This story continues at The Next Web