Yesterday I downloaded a local LLM app to my iPhone 15 Pro Phi3 mini was throwing 15 tokens / sec. Don't get sidetracked by Apple, everyone is going the same place, and very soon. 

 Which App?  

 This was https://github.com/guinmoon/LLMFarm 

 Thanks! You inspired me to search github for an android equivalent app. Found this one but I have not tested it, still downloading Q4_K_M 
https://github.com/nerve-sparks/iris_android
 

 search terms I used
https://github.com/search?q=ggml+android+&type=repositories