coreai-model-zoo
57 pre-converted Core AI models for Apple, with recipes and one-line Swift loading.
The definitive free resource for Apple developers wanting verifiable, reproducible on-device AI. If you're on iOS/macOS 27 and want local LLMs, this beats starting from scratch — you get 57 vetted models, conversion recipes, and one-line Swift loading. Strictly Apple-only, so cross-platform teams should look elsewhere.
Verified 7d ago · liveness 76/100 · cite: rightaichoice.com/tools/coreai-model-zoo
- iOS/macOS developers building on-device AI apps with Core AI
- AI engineers needing optimized conversion pipelines for Apple Silicon
- Developers targeting iPhone 17 Pro GPU/ANE for local inference
- Researchers exploring local LLM deployment on Apple devices
- Teams needing cross-platform model deployment (Android/Windows)
- Enterprises requiring commercial support or SLAs
- Developers without Apple devices running iOS 27 or macOS 27
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip CoreAI-Model-Zoo if you're not on Apple hardware running iOS 27 or macOS 27, or if you need models that aren't in the 57-model catalog and aren't willing to convert them yourself.
Large models are Mac-only; if you need them on iPhone, you'll have to find alternative smaller models or accept Mac-only deployment.
CoreAI-Model-Zoo is free and open-source, making it the most cost-effective option for Apple developers compared to paid services like Replicate or Hugging Face Inference Endpoints. There's no per-token or per-hour cost; you only pay for your own hardware.
In short
coreai-model-zoo — 57 pre-converted Core AI models for Apple, with recipes and one-line Swift loading. Best for iOS/macOS developers building on-device AI apps with Core AI, AI engineers needing optimized conversion pipelines for Apple Silicon, Developers targeting iPhone 17 Pro GPU/ANE for local inference. Free to use.
What people actually say about coreai-model-zoo — is it worth it?
We ran a structured research pass across product reviews, community discussions, and post-purchase forum threads to surface the patterns vendors won't publish themselves. Below: the recurring strengths, the hidden costs people mention most, and the cohort that consistently regrets adopting this tool.
22 mentions across 2 sources (Hacker News, YouTube) · researched Aug 15, 2026.
- +Free and open-source, with 57 pre-converted .aimodel files.
- +Covers a wide range of model types: LLMs, VLMs, OCR, ASR, TTS, and generation.
- +Reproducible conversions with recipes and zoo_convert.py for auditability.
- +One-line Swift loading via CoreAIKit is a real time-saver.
- +Custom Metal kernels for GPU/ANE acceleration are included.
- −No Windows or NVIDIA support; strictly Apple ecosystem only.
- −Performance on expensive Macs is questioned by some users.
- −No official support channels or documentation for common issues.
- −Requires advanced Swift and Core AI knowledge to use effectively.
- −Community feedback is sparse; few real-world validation reports.
- • Time to learn Core AI and integrate the zoo into existing projects
- • Potential cloud or infrastructure costs if you deploy at scale (but the zoo itself is local)
Viability Score
How well maintained and how widely used is coreai-model-zoo? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: August 2026
How we score →Key Features
- 57 pre-converted .aimodel files for LLM, VLM, OCR, ASR, TTS, image/video/music gen, forecasting
- Conversion recipes in recipe.toml, reproducible via zoo_convert.py run
- Verify bundles against source models with zoo_verify.py
- CoreAIKit Swift package for one-line model loading
- Integration with Apple FoundationModels API (tool calling, guided generation)
- Zero-dependency loading via CoreAILanguageModel(resourcesAt:)
- Custom Metal kernels for GPU/ANE acceleration
- iPhone 17 Pro GPU/ANE device-verified models
- Mac-only models clearly marked
- MoE and dense model support, including MLA (GLM-4.7-Flash)
- Official-QAT int4 quantization (Gemma 4)
- Vision-language models (Qwen3-VL, Holo2-4B, MiniCPM-V 4.6)
- Audio understanding (Qwen2.5-Omni-3B) and ASR (Whisper, Qwen3-ASR, Parakeet-TDT)
- Text-to-speech (Kokoro-82M, VoxCPM)
- Diffusion LLM (LLaDA-8B) and ternary VLA (BitVLA)
About coreai-model-zoo
CoreAI-Model-Zoo is a free, open-source GitHub repository for Apple developers who want to run AI models on-device without fighting conversion tools. It curates 57 pre-converted .aimodel files for Apple's Core AI framework on iOS 27 and macOS 27, spanning LLMs, vision-language models, OCR, ASR, TTS, image/video/music generation, and time-series forecasting. Each model is downloadable from Hugging Face, gated against its source model for parity, and ships with a recipe in models/<model>/recipe.toml that reproduces the conversion via zoo_convert.py. That means you're not just downloading weights — you get the exact pipeline that produced them, so you can rebuild, audit, or adapt any bundle yourself. The gateway to actually using these models is CoreAIKit, a Swift package that gives you one-line model loading. For example, `let chat = try await ChatSession(catalog: "qwen3.5-2b")` runs a chat model offline, cached after first download. The same gesture works across capabilities: `KitTranscriber(catalog: "whisper-large-v3-turbo")` is speech-to-text in three lines. Chat models also plug directly into Apple's FoundationModels API via `LanguageModelSession(model: try await KitLanguageModel(model: .qwen3_0_6B))`, giving you tool calling, @Generable guided generation, and transcripts. If you prefer zero dependencies, every bundle loads as-is with Apple's own `CoreAILanguageModel(resourcesAt:)`. Transparency is a core promise. Gates are run in the repository, and each model card documents what was measured on which hardware — iPhone-tier models are device-verified on iPhone 17 Pro GPU/ANE, while larger models are Mac-only and clearly marked. Where the shipped configuration could not be recovered from the repository, the recipe says so rather than guessing. The repo also includes custom Metal kernels for GPU/ANE acceleration, a knowledge base covering Core AI pitfalls and conversion gotchas, and community-contributed ports credited by name.
Behind the Verdict
CoreAI-Model-Zoo is a standout resource for Apple developers who want to run on-device AI without the usual pain of model conversion. The key differentiator is reproducibility: every model ships with the exact recipe that produced it, and you can rerun the conversion or verify the bundle against the source model yourself. This transparency is rare in the model zoo space and gives you real confidence in what you're deploying. The one-line Swift loading via CoreAIKit is a genuine convenience — you can go from clone to a working chat model in minutes, and the integration with Apple's FoundationModels API means you're not locked into a proprietary runtime. The breadth of 57 models across modalities (LLM, VLM, OCR, ASR, TTS, generation, forecasting) means most developers will find something useful without hunting elsewhere. Strengths: device-verified models on iPhone 17 Pro, custom Metal kernels for GPU/ANE, community ports with attribution, and a knowledge base that helps you avoid common Core AI pitfalls. Weaknesses: strictly Apple-only — iOS 27/macOS 27 required, and large models are Mac-only. The catalog is limited to 57 models; if you need something not listed, you'll have to convert it yourself using the provided tooling. There's no commercial support or SLA — this is a community project, so you're on your own if something breaks. Where it fits: indie developers, AI engineers, and researchers building on-device AI features for Apple platforms who value control and verifiability. Not for cross-platform teams or enterprises needing support.
Researching coreai-model-zoo? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Real-world workflow fit
Concrete scenarios for the personas coreai-model-zoo actually fits — and what changes day-one when you adopt it.
Building an offline translation app; needs on-device ASR and LLM.
Outcome: Clone coreai-kit, open ChatDemo, pick whisper-large-v3-turbo and a chat model, and integrate into the app within an hour, running fully offline.
Wants to prototype a privacy-preserving chatbot for a macOS app.
Outcome: Uses the pre-converted Qwen3.5-2B model with FoundationModels API integration to add tool calling and guided generation, with full control over the conversion recipe.
Evaluating small language models for edge deployment.
Outcome: Downloads multiple models, uses zoo_verify.py to confirm parity, and reads the knowledge base for conversion gotchas, saving weeks of trial and error.
Use Cases
- Run Qwen3.5-2B locally on iPhone 17 Pro for fast text generation without internet
- Deploy Gemma 4 E2B with int4 quantization on macOS 27 for privacy-sensitive apps
- Use GLM-4.7-Flash MoE model for multi-turn conversation on Apple Silicon
- Build a Swift-based chatbot using the provided runner and Core AI framework
- Study conversion gotchas to avoid common pitfalls when porting other LLMs to Apple Core AI
- Prototype on-device AI features for iOS 27 apps using pre-converted .aimodel files
Models Under the Hood
as of 2026-08-19
Limitations
- The model zoo requires iOS 27 or macOS 27 and is tested on iPhone 17 Pro.
- Models are pre-converted and downloadable from Hugging Face, but conversion scripts may need specific environment setup.
- No cloud API is provided; all inference is on-device.
- Large models are Mac-only, so you can't run them on iPhone.
- The catalog is limited to 57 models; if you need a model not listed, you'll have to convert it yourself using the provided tooling.
as of 2026-08-15
Verification history
We have re-verified coreai-model-zoo 6 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
Free to cite with attribution — this page re-verifies continuously.
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Plans compared
For each published coreai-model-zoo tier: who it actually fits, and what it adds vs. the previous tier. Cross-reference the cost calculator above for projected annual outlay.
Free
$0/mo
Ideal for
Apple developers and researchers who want a free, open-source collection of pre-converted on-device models with full transparency and reproducibility.
What this tier adds
Starting tier with $0/mo, unlimited access to all 57 models, recipes, CLI tools, and the CoreAIKit package — no paid upgrade path.
Where the pricing makes sense
The company stage and team size where coreai-model-zoo's pricing actually pencils out — and where peers do it cheaper.
CoreAI-Model-Zoo is free and open-source, making it the most cost-effective option for Apple developers compared to paid services like Replicate or Hugging Face Inference Endpoints. There's no per-token or per-hour cost; you only pay for your own hardware.
Setup time & first value
How long it actually takes to get something useful out of coreai-model-zoo — broken out by persona, not the marketing-page minute.
For a new user, expect under 15 minutes to clone the repo, open the ChatDemo Xcode project, and run your first model on a compatible device. If you need to convert a custom model, budget several hours to learn the tooling and troubleshoot environment specifics.
Switching to or from coreai-model-zoo
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From CoreML-Models to CoreAI-Model-Zoo: The zoo is the successor, offering pre-converted .aimodel bundles with recipes and a Swift package for one-line loading, so you can switch by replacing your CoreML models and
- →From manual model conversion with coremltools to CoreAI-Model-Zoo: Skip the conversion pipeline by using the pre-built bundles and recipes, and only convert custom models when needed.
- ↗To Hugging Face Hub: Download the original source models from Hugging Face and convert them yourself if you need versions not available in the zoo.
- ↗To a commercial on-device inference platform (e.g., Google's MediaPipe): You'll need to convert the models to a different format and adopt a different runtime, but the source models are open.
Integrations
Resources & Guides
- Resourcegithub.com
README · coreai-model-zoo
Helpful link from github.com
- Resourcegithub.com
CONTRIBUTING · coreai-model-zoo
Helpful link from github.com
- Resourcegithub.com
PORTING · coreai-model-zoo
Helpful link from github.com
- Resourcegithub.com
BENCHMARKS · coreai-model-zoo
Helpful link from github.com
- Resourcegithub.com
Knowledge · coreai-model-zoo
Helpful link from github.com
- Resourcegithub.com
Apps · coreai-model-zoo
Helpful link from github.com
Tutorials & Learning
Official links
Tools that pair well with coreai-model-zoo
Common stack mates teams adopt alongside coreai-model-zoo, with the specific reason each pairing earns its keep.
Featured Head-to-Head Comparisons
Coreai Model Zoo vs Voyage Ai
These tools serve completely different needs. Choose Voyage AI if you run an enterprise RAG pipeline needing domain-tuned embeddings and rerankers, especially for finance/legal; its 32K context and low-dim vectors reduce storage cost. Choose coreai-model-zoo if you're an Apple developer deploying LLMs on-device on iOS 27 / macOS 27 — it's free, open-source, and includes pre-converted models with GPU/ANE acceleration. No overlap in target audience.
Coreai Model Zoo vs Temporal Ai
These tools solve entirely different problems. Choose Temporal AI if you need a robust orchestration platform to build reliable, fault-tolerant AI agents and long-running workflows — it's overkill for simple tasks but essential for mission-critical systems. Choose CoreAI-Model-Zoo if you are an iOS/macOS developer targeting on-device AI with Apple's latest hardware and need ready-to-use, optimized models. There is no meaningful overlap; your choice depends on whether your bottleneck is workflow reliability or on-device model deployment.
Coreai Model Zoo vs Spider Cloud
Spider Cloud and CoreAI-Model-Zoo serve completely different needs: Spider Cloud is a cloud API for web data extraction (RAG, AI agents), while CoreAI-Model-Zoo is a free, on-device model repository for Apple developers. If you need live web data for AI pipelines at low cost, go with Spider Cloud. If you're building local AI apps on iPhone/Mac with optimized Core AI models, CoreAI-Model-Zoo is the obvious choice.
Alternatives to coreai-model-zoo
View all
Cortex.cpp
Run 123+ open-source models locally or connect online APIs in one free, open-source desktop app
Atomic Chat
Free, private, offline AI chat with 1000+ local LLMs, no account needed.
Frequently Asked Questions
Best-of guides
Used coreai-model-zoo? Help shape our editorial sentiment research.