Link CopyTwitterLinkedinWhatsappApple Is Testing Tech That Shrinks a 54GB AI Model to 3. 9GB—and Fits It on a PhoneA Caltech spinout says it fits a 54 GB model into 3.9 GB on a phone, and Apple started testing it the week its Siri beta went public. The timing is the story.Apple is in early talks with PrismML, a Caltech spinout that says it compressed Alibaba's 27-billion-parameter Qwen 3.6 model from roughly 54 GB to 3.9 GB, bringing what the company calls the first 27B-class model onto a phone. PrismML chief executive Babak Hassibi told CNBC that Apple and other companies have been evaluating the startup's models for speed, energy efficiency and on-device performance. "They're really evaluating our technology right now," he said of Apple, in remarks carried by AppleInsider, describing the discussions as very early and their destination as still open. The confirmation landed one day after Apple opened the iOS 27 public beta, giving iPhone owners their first broad access to the long-delayed Siri overhaul, and the collision of those two dates explains most of what follows. Apple already ships a 20-billion-parameter model on its priciest iPhones; PrismML is pitching a bigger one that stays switched on all the way through.The Claim, In NumbersAbout The AuthorAt heart, I am a storyteller drawn to the watershed moments that bend the technology landscape. I braid narrative with data, humanise statistics, and trace the arc from first spark to world-changing impact. My reportage, features and reviews are witty, sardonic, visual and vivid, using anecdote to illuminate rather than eviscerate.
Apple Is Testing Tech That Shrinks a 54GB AI Model to 3. 9GB—and Fits It on a Phone
A Caltech spinout says it fits a 54 GB model into 3.9 GB on a phone, and Apple started testing it the week its Siri beta went public. The timing is the story.
4,047 words~18 min read
