Apple has their own models under the hood I believe. I remember from like a year or two ago they had an open line called "ELM" (Efficient Language Model), but I'm not sure if that's what they're actually using.
I am excited to see what the benchmarks look like though, once it's live.
They described their home-grown models last year: https://machinelearning.apple.com/research/introducing-apple...
I'm assuming this is an updated version of those.
Apple has their own models under the hood I believe. I remember from like a year or two ago they had an open line called "ELM" (Efficient Language Model), but I'm not sure if that's what they're actually using.
I am excited to see what the benchmarks look like though, once it's live.
they also use their ANE and CoreML for smaller on-device stuff
https://huggingface.co/apple
They're also working with Anthropic on a coding platform: https://www.macrumors.com/2025/05/02/apple-anthropic-ai-codi...