Bonsai 27B

Bonsai 27B is a large language model with 27 billion parameters, designed specifically to operate on smartphones. This breakthrough addresses the common limitation of running powerful AI models locally on mobile hardware, which typically requires cloud connectivity or heavy computational resources. By enabling on-device inference, Bonsai 27B offers privacy benefits, offline functionality, and reduced latency for users. The model was announced via PrismML, suggesting a focus on practical deployment in consumer applications. While details on architecture and training data are still emerging, the core value proposition is clear: bringing desktop-class AI capabilities to mobile devices without sacrificing performance.

Bonsai 27B gained attention after a launch announcement on PrismML, highlighting its ability to run a 27B model on a phone, which is a significant technical achievement in on-device AI.