Comment by linzhangrun

Another thing these AI giants should worry about is the local deployment

A $4,699 DGX Spark can easily run Qwen 3.6 35B, and its performance crushes GPT-o1, the SOTA from two years ago