Apple is reportedly developing a large language model (LLM) called ReALM to boost Siri’s capabilities and improve contextual understanding. This model, available in four variants, is claimed to perform on par with OpenAI’s ChatGPT, according to an Apple research paper shared by Windows Central.
ReALM is designed to enhance Siri’s conversational abilities by understanding context and processing onscreen content, giving it an edge over ChatGPT. Apple’s ReALM is set to launch in four sizes: ReALM-80M, ReALM-250M, ReALM-1B, and ReALM-3B, indicating varying levels of processing power.
Comparisons between ReALM and GPT-3.5 and GPT-4 show that even the smallest ReALM-80M model is on par with GPT-4, with larger models performing significantly better. ReALM demonstrates significant improvements in processing onscreen information, achieving up to 5% better results.
While Apple has been relatively quiet about AI developments compared to Microsoft, rumors suggest that ReALM could be Apple’s entrance into the AI landscape, with plans to enhance Siri’s AI capabilities. The launch of ReALM may be discussed at WWDC in June.
Editors’ Recommendations
