PrismML says its compressed version of Alibaba’s Qwen model uses up to 15 times less memory, potentially advancing Apple’s AI push.