PrismML's Compressed AI Model Brings On-Device Intelligence to Apple Devices
A California-based startup called PrismML has made significant strides in the field of artificial intelligence by developing a compressed AI model that can run locally on Apple's iPhone 17 Pro without requiring a cloud connection.
The model, called Bonsai 27B, was released on July 14 and is built on Alibaba's open-weight Qwen3.6 architecture. Despite being compressed to just 3.9 gigabytes, it retains roughly 90% of the original model's performance, making it a meaningful technical achievement in the field of model compression.
Bonsai 27B is not only compatible with Apple devices such as the iPhone 17 Pro but also with iPad, Mac, and PC devices, broadening its potential reach across Apple's entire hardware ecosystem. The startup's CEO recently told CNBC that Apple is in 'very early' discussions about the technology.
The emergence of capable, offline AI like Bonsai 27B may signal an era where consumers can enjoy AI without relying on monthly subscriptions or cloud services. This could position Apple well if the broader market shifts toward locally run, open-weight AI models.