Apple Announces Foundation Models Framework for Developers to Leverage AI

Aren't the on device models quite limited in capabilities? What can they do? In any case even access to a limited model could be huge.

While Apple’s OpenELM LLM is no ChatGPT, it is a very capable resource for incorporating conversational interface, on-device RAG and LoRA fine-tuning into an app.

We used it to incorporate a fine-tuned model into one of our apps and was very pleased with the results. We held off on shipping the updated app because at the time it would require the app to download a 3.8GB fine-tuned OpenELM model from huggingface — which we didn’t want to require users to do.

We also tried implementing the same AI feature set by incorporating ChatGPT and Perplexity via API. It worked functionally but the latency and API costs were prohibitive for our use case.

Need to see the details but this announcement could possibly eliminate all of these problems — assuming the foundation models include OpenELM equivalents and the API supports fine-tuning via Adapters as was announced at last year’s WWDC. I can’t wait to see the details and start working with the beta!

Follow-up: After watching the WWDC Platforms State of Union keynote .. AFM Framework is a Bonafide Game Changer!

The AFM framework and APIs exceed my expectations by delivering utility that goes well beyond providing the desired functionality — which for me was simply (a) a capable on-device LLM that eliminates the need for costly, high-latency, off-board 3rd-party LLMs, (b) adapters to allow model fine-tuning, (c) user data privacy, and (d) off-line operation.

The unexpected benefits include tool calling, response streaming and model macros that eliminate complex and error-prone LLM response parsing and mapping to app data structures.

I took the last item (complex, imprecise and time-consuming parsing and mapping) as a given and something that developers should just expect to do when incorporating LLMs into an app with structured data — and was completely surprised to hear that Apple completely eliminated this issue for Apple platform developers! This is a really big deal because this single issue is actually a limiting factor for app development use cases. Prior to the Apple specialized macro utility the solution was either complex and brittle regular expressions that were guaranteed to fail (because LLM output is non-deterministic), or ballooning LLM API cost and latency to add guardrails to constrain LLM output to behave more deterministically.

The final word will depend on the stability of the AFM implementation and how well it aligns with what was demonstrated, but this developer is very pleased. The AFM API is a year late, but definitely way better than what was expected.

Bravo Apple. Thank you! ??

This is huge. I have so many ideas for my app now! I can't wait.

I think this could open up a lot of new and exciting apps for developers to build, but probably at the expense of battery life. It will be interesting to see how this evolves. I think this will be one of those things that is a much bigger deal a few years down the road, so better to see it now than later.

('https://www.macrumors.com/2025/06/09/foundation-models-framework/')

Apple at WWDC today announced Foundation Models Framework, a new API allowing third-party developers to leverage the large language models at the heart of Apple Intelligence and build it into their apps.



With the Foundation Models Framework, developers can integrate Apple's on-device models directly into apps, allowing them to build on Apple Intelligence.
The Foundation Models framework lets developers build AI-powered features that work offline, protect privacy, and incur no inference costs. For example, an education app can generate quizzes from user notes on-device, and an outdoors app can offer offline natural language search.

Apple says the framework is available for testing starting today through the Apple Developer Program at developer.apple.com, and a public beta will be available through the Apple Beta Software Program next month at beta.apple.com. It includes built-in features like guided generation and tool calling for easy integration of generative capabilities into existing apps.

Article Link: Apple Announces Foundation Models Framework for Developers to Leverage AI ('https://www.macrumors.com/2025/06/09/foundation-models-framework/')

Finally!! Been waiting a year for this. Need to see the details but this could be a game changer for devs.

Aren't the on device models quite limited in capabilities? What can they do? In any case even access to a limited model could be huge.

They have two main capabilities
- "understanding" language and
- "understanding" images.

The obvious thing you can do is baby steps towards a language-driven UI. Imagine something like you tell UberEats "What was that Asian food I ordered last week? Can you order me that again?"
I think at least part of why Apple is doing this is research, to see how this plays out in the real world.

There are also some less obvious capabilities this allows. For example imagine a note-taking app that creates quizzes from your last week of notes, so you can see what you remember, vs what you don't remember. (I used the word "remember" here deliberately. A better app would create quizzes to see what you UNDERSTAND, but that's probably still too much to expect, even from a leading edge LLM, let along a small edge model.)

Similarly presumably we will see things like photo editing apps where you can just tell the app "remove Jenna's face" and see what happens.
Again this is research. Ultimately the goal is a system-wide language UI, not dedicated per-app code handling this stuff. But at least this gets Apple some of the way there for a year or two, while they figure out the bigger solution.