I am writing a custom package wrapping Foundation Models which provides a chain-of-thought with intermittent self-evaluation among other things. At first I was designing this package with the command line in mind, but after seeing how well it augments the models and makes them more intelligent I wanted to try and build a SwiftUI wrapper around the package. When I started I was using synchronous generation rather than streaming, but to give the best user experience (as I've seen in the WWDC sessions) it is necessary to provide constant feedback to the user that something is happening. I have created a super simplified example of my setup so it's easier to understand. First, there is the Reasoning conversation item, which can be converted to an XML representation which is then fed back into the model (I've found XML works best for structured input) public typealias ConversationContext = XMLDocument extension ConversationContext { public func toPlainText() -> String { return xmlString(options: [.nodePrettyPrint]) } } /// Represents a reasoning item in a conversation, which includes a title and reasoning content. /// Reasoning items are used to provide detailed explanations or justifications for certain decisions or responses within a conversation. @Generable(description: "A reasoning item in a conversation, containing content and a title.") struct ConversationReasoningItem: ConversationItem { @Guide(description: "The content of the reasoning item, which is your thinking process or explanation") public var reasoningContent: String @Guide(description: "A short summary of the reasoning content, digestible in an interface.") public var title: String @Guide(description: "Indicates whether reasoning is complete") public var done: Bool } extension ConversationReasoningItem: ConversationContextProvider { public func toContext() -> ConversationContext { // <ReasoningItem title="${title}"> // ${reasoningContent} // </ReasoningItem> let root = XMLElement(name: "ReasoningItem") root.addAttribute(XMLNode.attribute(withName: "title", stringValue: title) as! XMLNode) root.stringValue = reasoningContent return ConversationContext(rootElement: root) } } Then there is the generator, which creates a reasoning item from a user query and previously generated items: struct ReasoningItemGenerator { var instructions: String { """ <omitted for brevity> """ } func generate(from input: (String, [ConversationReasoningItem])) async throws -> sending LanguageModelSession.ResponseStream<ConversationReasoningItem> { let session = LanguageModelSession(instructions: instructions) // build the context for the reasoning item out of the user's query and the previous reasoning items let userQuery = "User's query: \(input.0)" let reasoningItemsText = input.1.map { $0.toContext().toPlainText() }.joined(separator: "\n") let context = userQuery + "\n" + reasoningItemsText let reasoningItemResponse = try await session.streamResponse( to: context, generating: ConversationReasoningItem.self) return reasoningItemResponse } } I'm not sure if returning LanguageModelSession.ResponseStream<ConversationReasoningItem> is the right move, I am just trying to imitate what session.streamResponse returns. Then there is the orchestrator, which I can't figure out. It receives the streamed ConversationReasoningItems from the Generator and is responsible for streaming those to SwiftUI later and also for evaluating each reasoning item after it is complete to see if it needs to be regenerated (to keep the model on-track). I want the users of the orchestrator to receive partially generated reasoning items as they are being generated by the generator. Later, when they finish, if the evaluation passes, the item is kept, but if it fails, the reasoning item should be removed from the stream before a new one is generated. So in-flight reasoning items should be outputted aggresively. I really am having trouble figuring this out so if someone with more knowledge about asynchronous stuff in Swift, or- even better- someone who has worked on the Foundation Models framework could point me in the right direction, that would be awesome!

I'm testing Foundation Model on my iPad Pro (5th gen) iOS 26. Up until late this morning, I can no longer load the SystemLanguageModel.default. I'm not doing anything interesting, something as basic as this is only going to unavailable, specifically I get unavailable reason: modelNotReady. let model = SystemLanguageModel.default ... switch model.availability { case .available: print("LM available") case .unavailable(let reason): print("unavailable reason: ", String(describing: reason)) } I also ran the FoundationModelsTripPlanner app, same thing. It was working yesterday, I have not modified that project either. Why is the Model not ready? How do I fix this? Yes, I tried restarting both my laptop and iPad, no luck.

I've spent way too long today trying to convert an Object Detection TensorFlow2 model to a CoreML object classifier (with bounding boxes, labels and probability score) The 'SSD MobileNet v2 320x320' is here: https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/tf2_detection_zoo.md And I've been following all sorts of posts and ChatGPT https://apple.github.io/coremltools/docs-guides/source/tensorflow-2.html#convert-a-tensorflow-concrete-function https://developer.apple.com/videos/play/wwdc2020/10153/?time=402 To convert it. I keep hitting the same errors though, mostly around: NotImplementedError: Expected model format: [SavedModel | concrete_function | tf.keras.Model | .h5 | GraphDef], got <ConcreteFunction signature_wrapper(input_tensor) at 0x366B87790> I've had varying success including missing output labels/predictions. But I simply want to create the CoreML model with all the right inputs and outputs (including correct names) as detailed in the docs here: https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/running_on_mobile_tf2.md It goes without saying I don't have much (any) experience with this stuff including Python so the whole thing's been a bit of a headache. If anyone is able to help that would be great. FWIW I'm not attached to any one specific model, but what I do need at minimum is a CoreML model that can detect objects (has to at least include lights and lamps) within a live video image, detecting where in the image the object is. The simplest script I have looks like this: import coremltools as ct import tensorflow as tf model = tf.saved_model.load("~/tf_models/ssd_mobilenet_v2_320x320_coco17_tpu-8/saved_model") concrete_func = model.signatures[tf.saved_model.DEFAULT_SERVING_SIGNATURE_DEF_KEY] mlmodel = ct.convert( concrete_func, source="tensorflow", inputs=[ct.TensorType(shape=(1, 320, 320, 3))] ) mlmodel.save("YourModel.mlpackage", save_format="mlpackage")

I'm new to Swift and was hoping the Playground would support loading adaptors. When I tried, I got a permissions error - thinking it's because it's not in the project and Playgrounds don't like going outside the project? A tutorial and some sample code would be helpful. Also some benchmarks on how long it's expected to take. Selfishly I'm on an M2 Mac Mini.

Download the Foundation Models Adaptor Training Toolkit Hi, after I clicked on the download button, I was redirected to this page https://developer.apple.com and did not download the toolkit.

When I initialize a session with an existing transcript using this initializer: public convenience init(model: SystemLanguageModel = .default, guardrails: LanguageModelSession.Guardrails = .default, tools: [any Tool] = [], transcript: Transcript) The tools get ignored. I noticed that when doing that, the model never use the tools. When inspecting the transcript, I can see that the instruction entry does not have any tools available to it. I tried this for both transcripts that already include an instruction entry and ones that don't - both yielding the same result.. Is this the intended behavior / am I missing something here?

Hey everyone I'm Manish Mehta, field CTO at Centific. I recently read Apple's white paper, The Illusion of Thinking and it got me thinking about the current state of AI reasoning. Who here has read it? The paper highlights how LLMs often rely on pattern recognition rather than genuine understanding. When faced with complex tasks, their performance can degrade significantly. I was just thinking that to move beyond this problem, we need to explore approaches that combines Deeper Reasoning Architectures for true cognitive capability with Deep Human Partnership to guide AI toward better judgment and understanding. The first part means fundamentally rewiring AI to reason. This involves advancing deeper architectures like World Models, which can build internal simulations to understand real-world scenarios , and Neurosymbolic systems, which combines neural networks with symbolic reasoning for deeper self-verification. Additionally, we need to look at deep human partnership and scalable oversight. An AI cannot learn certain things from data alone, it lacks the real-world judgment an AI will never have. Among other things, deep domain expert human partners are needed to instill this wisdom , validate the AI's entire reasoning process , build its ethical guardrails , and act as skilled adversaries to find hidden flaws before they can cause harm. What do you all think? Is this focus on a deeper partnership between advanced AI reasoning and deep human judgment the right path forward? Agree? Disagree? Thanks

I am using gemini2.5-flash with SwiftUI. How can I receive a response in JSON?

With respond() methods, the foundation model works well enough. With streamResponse() methods, the responses are very repetitive, verbose, and messy. My app with foundation model uses more than 500 MB memory on an iPad Pro when running from Xcode. Devices supporting Apple Intelligence have at least 8GB memory. Should Apple use a bigger model (using 3 ~ 4 GB memory) for better stream responses?

I am using Foundation Models for the first time and no response is being provided to me. Code import Playgrounds import FoundationModels #Playground { let session = LanguageModelSession() let result = try await session.respond(to: "List all the states in the USA") print(result.content) } Canvas Output What I did New file Code Canvas refreshes but nothing happens Am I missing a step or setup here? Please help. Something so basic is not working I do not know what to do. Running 40GPU, 16CPU MacBook Pro.. IOS26/Xcodebeta2/Tahoe allocated 8CPU, 48GB memory in Parallels VM. Settings for Playgrounds in Xcode Thank you for your help in advance.

I am using a contact tool to help get contact from my address book. but the model ins't invoking my tool call method. Even tried with a simple tool the outcome is the same my simple tool is not being invoked.

Hello Team, I'm currently working on a proof of concept using Apple's Foundation Model for a RAG-based chat system on my MacBook Pro with the M1 Max chip. Environment details: macOS: 26.0 Beta Xcode: 26.0 beta 2 (17A5241o) Target platform: iPad (as the iPhone simulator does not support Foundation models) While testing, even with very small input prompts to the LLM, I intermittently encounter the following error: InferenceError::inference-Failed::Failed to run inference: Context length of 4096 was exceeded during singleExtend. Has anyone else experienced this issue? Are there known limitations or workarounds for context length handling in this setup? Any insights would be appreciated. Thank you!

I'm experimenting with downloading an audio file of spoken content, using the Speech framework to transcribe it, then using FoundationModels to clean up the formatting to add paragraph breaks and such. I have this code to do that cleanup: private func cleanupText(_ text: String) async throws -> String? { print("Cleaning up text of length \(text.count)...") let session = LanguageModelSession(instructions: "The content you read is a transcription of a speech. Separate it into paragraphs by adding newlines. Do not modify the content - only add newlines.") let response = try await session.respond(to: .init(text), generating: String.self) return response.content } The content length is about 29,000 characters. And I get this error: InferenceError::inferenceFailed::Failed to run inference: Context length of 4096 was exceeded during singleExtend.. Is 4096 a reference to a max input length? Or is this a bug? This is running on an M1 iPad Air, with iPadOS 26 Seed 1.

At WWDC25 we launched a new type of Lab event for the developer community - Group Labs. A Group Lab is a panel Q&A designed for a large audience of developers. Group Labs are a unique opportunity for the community to submit questions directly to a panel of Apple engineers and designers. Here are the highlights from the WWDC25 Group Lab for Machine Learning and AI Frameworks. What are you most excited about in the Foundation Models framework? The Foundation Models framework provides access to an on-device Large Language Model (LLM), enabling entirely on-device processing for intelligent features. This allows you to build features such as personalized search suggestions and dynamic NPC generation in games. The combination of guided generation and streaming capabilities is particularly exciting for creating delightful animations and features with reliable output. The seamless integration with SwiftUI and the new design material Liquid Glass is also a major advantage. When should I still bring my own LLM via CoreML? It's generally recommended to first explore Apple's built-in system models and APIs, including the Foundation Models framework, as they are highly optimized for Apple devices and cover a wide range of use cases. However, Core ML is still valuable if you need more control or choice over the specific model being deployed, such as customizing existing system models or augmenting prompts. Core ML provides the tools to get these models on-device, but you are responsible for model distribution and updates. Should I migrate PyTorch code to MLX? MLX is an open-source, general-purpose machine learning framework designed for Apple Silicon from the ground up. It offers a familiar API, similar to PyTorch, and supports C, C++, Python, and Swift. MLX emphasizes unified memory, a key feature of Apple Silicon hardware, which can improve performance. It's recommended to try MLX and see if its programming model and features better suit your application's needs. MLX shines when working with state-of-the-art, larger models. Can I test Foundation Models in Xcode simulator or device? Yes, you can use the Xcode simulator to test Foundation Models use cases. However, your Mac must be running macOS Tahoe. You can test on a physical iPhone running iOS 18 by connecting it to your Mac and running Playgrounds or live previews directly on the device. Which on-device models will be supported? any open source models? The Foundation Models framework currently supports Apple's first-party models only. This allows for platform-wide optimizations, improving battery life and reducing latency. While Core ML can be used to integrate open-source models, it's generally recommended to first explore the built-in system models and APIs provided by Apple, including those in the Vision, Natural Language, and Speech frameworks, as they are highly optimized for Apple devices. For frontier models, MLX can run very large models. How often will the Foundational Model be updated? How do we test for stability when the model is updated? The Foundation Model will be updated in sync with operating system updates. You can test your app against new model versions during the beta period by downloading the beta OS and running your app. It is highly recommended to create an "eval set" of golden prompts and responses to evaluate the performance of your features as the model changes or as you tweak your prompts. Report any unsatisfactory or satisfactory cases using Feedback Assistant. Which on-device model/API can I use to extract text data from images such as: nutrition labels, ingredient lists, cashier receipts, etc? Thank you. The Vision framework offers the RecognizeDocumentRequest which is specifically designed for these use cases. It not only recognizes text in images but also provides the structure of the document, such as rows in a receipt or the layout of a nutrition label. It can also identify data like phone numbers, addresses, and prices. What is the context window for the model? What are max tokens in and max tokens out? The context window for the Foundation Model is 4,096 tokens. The split between input and output tokens is flexible. For example, if you input 4,000 tokens, you'll have 96 tokens remaining for the output. The API takes in text, converting it to tokens under the hood. When estimating token count, a good rule of thumb is 3-4 characters per token for languages like English, and 1 character per token for languages like Japanese or Chinese. Handle potential errors gracefully by asking for shorter prompts or starting a new session if the token limit is exceeded. Is there a rate limit for Foundation Models API that is limited by power or temperature condition on the iPhone? Yes, there are rate limits, particularly when your app is in the background. A budget is allocated for background app usage, but exceeding it will result in rate-limiting errors. In the foreground, there is no rate limit unless the device is under heavy load (e.g., camera open, game mode). The system dynamically balances performance, battery life, and thermal conditions, which can affect the token throughput. Use appropriate quality of service settings for your tasks (e.g., background priority for background work) to help the system manage resources effectively. Do the foundation models support languages other than English? Yes, the on-device Foundation Model is multilingual and supports all languages supported by Apple Intelligence. To get the model to output in a specific language, prompt it with instructions indicating the user's preferred language using the locale API (e.g., "The user's preferred language is en-US"). Putting the instructions in English, but then putting the user prompt in the desired output language is a recommended practice. Are larger server-based models available through Foundation Models? No, the Foundation Models API currently only provides access to the on-device Large Language Model at the core of Apple Intelligence. It does not support server-side models. On-device models are preferred for privacy and for performance reasons. Is it possible to run Retrieval-Augmented Generation (RAG) using the Foundation Models framework? Yes, it is possible to run RAG on-device, but the Foundation Models framework does not include a built-in embedding model. You'll need to use a separate database to store vectors and implement nearest neighbor or cosine distance searches. The Natural Language framework offers simple word and sentence embeddings that can be used. Consider using a combination of Foundation Models and Core ML, using Core ML for your embedding model.

I’ve been testing silent Siri engagement via typing on iOS 18 and also on iOS 26 beta 1 and beta 2. While normal typing works perfectly in type-to-Siri mode, I’ve noticed that swipe-to-type gestures don’t work within Siri’s input field. Interestingly, you still feel the usual haptic feedback associated with swipe typing, but no text appears in the Siri text box. Swipe-to-type continues to work flawlessly in other apps like Messages and Notes, so this seems to be an issue specific to Siri’s typing input handler in these betas. Hopefully, it will be fixed in the next release because swipe typing is essential to my silent Siri workflow.

When I try to run visionOS 26 beta 2 on my device the app crashes on Launch: dyld[904]: Symbol not found: _$s16FoundationModels10TranscriptV7entriesACSayAC5EntryOG_tcfC Referenced from: <A71932DD-53EB-39E2-9733-32E9D961D186> /private/var/containers/Bundle/Application/53866099-99B1-4BBD-8C94-CD022646EB5D/VisionPets.app/VisionPets.debug.dylib Expected in: <F68A7984-6B48-3958-A48D-E9F541868C62> /System/Library/Frameworks/FoundationModels.framework/FoundationModels Symbol not found: _$s16FoundationModels10TranscriptV7entriesACSayAC5EntryOG_tcfC Referenced from: <A71932DD-53EB-39E2-9733-32E9D961D186> /private/var/containers/Bundle/Application/53866099-99B1-4BBD-8C94-CD022646EB5D/VisionPets.app/VisionPets.debug.dylib Expected in: <F68A7984-6B48-3958-A48D-E9F541868C62> /System/Library/Frameworks/FoundationModels.framework/FoundationModels dyld config: DYLD_LIBRARY_PATH=/usr/lib/system/introspection DYLD_INSERT_LIBRARIES=/usr/lib/libLogRedirect.dylib:/usr/lib/libBacktraceRecording.dylib:/usr/lib/libMainThreadChecker.dylib:/usr/lib/libViewDebuggerSupport.dylib:/System/Library/PrivateFrameworks/GPUToolsCapture.framework/GPUToolsCapture Symbol not found: _$s16FoundationModels10TranscriptV7entriesACSayAC5EntryOG_tcfC Referenced from: <A71932DD-53EB-39E2-9733-32E9D961D186> /private/var/containers/Bundle/Application/53866099-99B1-4BBD-8C94-CD022646EB5D/VisionPets.app/VisionPets.debug.dylib Expected in: <F68A7984-6B48-3958-A48D-E9F541868C62> /System/Library/Frameworks/FoundationModels.framework/FoundationModels dyld config: DYLD_LIBRARY_PATH=/usr/lib/system/introspection DYLD_INSERT_LIBRARIES=/usr/lib/libLogRedirect.dylib:/usr/lib/libBacktraceRecording.dylib:/usr/lib/libMainThreadChecker.dylib:/usr/lib/libViewDebuggerSupport.dylib:/System/Library/PrivateFrameworks/GPUToolsCapture.framework/GPUToolsCapture Message from debugger: Terminated due to signal 6

For example: I have a list of to-dos, each with a unique id (a GUID). I want to feed them to the LLM model and have the model rewrite the items so they start with an action verb. I'd like to get them back and identify which rewritten item corresponds to which original item. I obviously can't compare the text, as it has changed. I've tried passing the original GUIDs in with each to-do, but the extra GUID characters pollutes the input and confuses the model. I've tried numbering them in order and adding an originalSortOrder field to my generable type, but it doesn't work reliably. Any suggestions? I could do them one at a time, but I also have a use case where I'm asking for them to be organized in sections, and while I've instructed the model not to rename anything, it still happens. It's just all very nondeterministic.

Hi, I have trained a basic adapter using the adapter training toolkit. I am trying a very basic example of loading it and running inference with it, but am getting the following error: Passing along InferenceError::inferenceFailed::loadFailed::Error Domain=com.apple.TokenGenerationInference.E5Runner Code=0 "Failed to load model: ANE adapted model load failure: createProgramInstanceWithWeights:modelToken:qos:baseModelIdentifier:owningPid:numWeightFiles:error:: Program load new instance failure (0x170006)." UserInfo={NSLocalizedDescription=Failed to load model: ANE adapted model load failure: createProgramInstanceWithWeights:modelToken:qos:baseModelIdentifier:owningPid:numWeightFiles:error:: Program load new instance failure (0x170006).} in response to ExecuteRequest Any ideas / direction? For testing I am including the .fmadapter file inside the app bundle. This is where I load it: @State private var session: LanguageModelSession? // = LanguageModelSession() func loadAdapter() async throws { if let assetURL = Bundle.main.url(forResource: "qasc---afm---4-epochs-adapter", withExtension: "fmadapter") { print("Asset URL: \(assetURL)") let adapter = try SystemLanguageModel.Adapter(fileURL: assetURL) let adaptedModel = SystemLanguageModel(adapter: adapter) session = LanguageModelSession(model: adaptedModel) print("Loaded adapter and updated session") } else { print("Asset not found in the main bundle.") } } This seems to work fine as I get to the log Loaded adapter and updated session. However when the below inference code runs I get the aforementioned error: func sendMessage(_ msg: String) { self.loading = true if let session = session { Task { do { let modelResponse = try await session.respond(to: msg) DispatchQueue.main.async { self.response = modelResponse.content self.loading = false } } catch { print("Error: \(error)") DispatchQueue.main.async { self.loading = false } } } } }

I am calling into an app extension from a Safari Web Extension (sendNativeMessage, which in turn results in a call to NSExtensionRequestHandling’s beginRequest). My Safari extension aims to make use of the new foundation models for some of the features it provides. In my testing, I hit the rate limit by sending 4 requests, waiting 30 seconds between each. This makes the FoundationModels framework (which would otherwise serve my use case perfectly well) unusable in this context, because the model is called in response to user input, and this rate of user input is perfectly plausible in a real world scenario. The error thrown as a result of the rate limit is “Safety guardrail was triggered after consecutive failures during streaming.", but looking at the system logs in Console.app shows the rate limit as the real culprit. My suggestions: Please introduce sensible rate limits for app extensions, through an entitlement if need be. If it is rate limited to 1 request per every couple of seconds, that would already fix the issue for me. Please document the rate limit. Please make the thrown error reflect that it is the result of a rate limit and not a generic guardrail violation. IMPORTANT: please indicate in the thrown error when it is safe to try again. Filed a feedback here: FB18332004

Greetings! I was trying to get a response from the LanguageModelSession but I just keep getting the following: Error getting response: Model Catalog error: Error Domain=com.apple.UnifiedAssetFramework Code=5000 "There are no underlying assets (neither atomic instance nor asset roots) for consistency token for asset set com.apple.MobileAsset.UAF.FM.Overrides" UserInfo={NSLocalizedFailureReason=There are no underlying assets (neither atomic instance nor asset roots) for consistency token for asset set com.apple.MobileAsset.UAF.FM.Overrides} This occurs both in macOS 15.5 running the new Xcode beta with an iOS 26 simulator, and also on a macOS 26 with Xcode beta. The simulators are both Pro iPhone 16s. I was wondering if anyone had any advice?