I am signed in into my paid ChatGPT account in Settings.
To test this thing out, I asked Apple Intelligence via Writing Tools to create an article about Vision Pro and Vision OS, avoiding specifically mentioning anything about Apple.
And as a result, I got some super generic response as if the LLM made a prediction on what these "technologies" would be in the LLM's opinion.
Then I asked the very same question in ChatGPT's Mac app using GPT-4o model, and got a very precise and well-structured answer about Apple's Vision Pro and Vision OS.
Which got me thinking: does Apple Intelligence use the outdated GPT-4 model rather than 4o?
To test this thing out, I asked Apple Intelligence via Writing Tools to create an article about Vision Pro and Vision OS, avoiding specifically mentioning anything about Apple.
And as a result, I got some super generic response as if the LLM made a prediction on what these "technologies" would be in the LLM's opinion.
Then I asked the very same question in ChatGPT's Mac app using GPT-4o model, and got a very precise and well-structured answer about Apple's Vision Pro and Vision OS.
Which got me thinking: does Apple Intelligence use the outdated GPT-4 model rather than 4o?