Apple has released OpenELM, a series of open-source small language models, for device usage. The models are specifically designed for text-related tasks such as writing emails.

OpenELM consists of a 270 million parameter model, a 450 million parameter model, a 1.1 billion parameter model, and a 3 billion parameter model. Pre-trained and instruction-tuned models are available for each parameter, offering eight types of models in total.

Apple claims that the models are cost-effective and are optimized to function effectively on devices such as smartphones and laptops. 

Analyst QuickTake : While this may indicate plans to incorporate GenAI models into its smartphones, the company is reportedly in talks with Google and Baidu to incorporate their models into the iPhone. Last month, Apple launched MM1, a multimodal language learning mode , and introduced ReALM (Reference Resolution as Language Modeling), a new AI system to improve voice assistants' interpretation and response to user commands this month.