coreai-model-zoo
FreeCurated Apple Core AI model zoo for iOS/macOS 27 with 30+ verified models
By Tanmay Verma, Founder · Last verified 05 Jul 2026
1 views
Added 6d ago
69/100MonitorIn short
coreai-model-zoo — Curated Apple Core AI model zoo for iOS/macOS 27 with 30+ verified models. 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.
Is coreai-model-zoo actually worth it?
LiveSee what real users actually say. We scan live discussions, reviews and complaints across the web and hand you an honest verdict — in under a minute.
3 free scans · no card needed · downloadable report
Editorial Verdict
Best for
iOS/macOS developers building on-device AI apps with Core AIAI engineers needing optimized conversion pipelines for Apple SiliconDevelopers targeting iPhone 17 Pro GPU/ANE for local inferenceResearchers exploring local LLM deployment on Apple devicesHobbyists wanting to run open-source LLMs on personal Apple hardware
Not ideal for
Teams needing cross-platform model deployment (Android/Windows)Enterprises requiring commercial support or SLAsDevelopers without Apple devices running iOS 27 or macOS 27Beginners new to machine learning or Core AI framework
The most practical resource for Apple developers deploying LLMs on-device. Unique value in pre-converted models with verified performance and conversion gotchas. However, it's strictly Apple-only and requires familiarity with Core AI.
Skip coreai-model-zoo if Skip CoreAI-Model-Zoo if you need cross-platform model deployment, commercial support, or have no Apple device running iOS 27/macOS 27.
Compare with: coreai-model-zoo vs Cortex.cpp, coreai-model-zoo vs Ollama, coreai-model-zoo vs Cohere
Last verified: July 2026
What independent users actually report about coreai-model-zoo
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.
24 mentions across 3 sources (Hacker News, YouTube, GitHub).
53% positive47% critical
Recurring strengths
- +Pre-converted .aimodel files for popular models like Qwen3.5 and Gemma 4.
- +End-to-end conversion scripts with detailed gotcha documentation.
- +Custom Metal kernels for GPU/ANE acceleration verified on iPhone 17 Pro.
- +Swift runner provided for easy on-device inference.
- +Supports modern architectures: MoE, dense, MLA, and int4 quantization.
Recurring frustrations
- −Very little community feedback to validate reliability.
- −No official support, issue responses, or maintenance guarantees.
- −Verified only on iPhone 17 Pro, an unreleased device.
- −Tied to iOS 27/macOS 27 — not backward compatible.
- −YouTube comments are off-topic; no organic user discourse.
Patterns worth knowing
Model conversion and deployment ease is the key draw
Seen on GitHub, Hacker News
Community is nascent — thin but appreciative
Seen on GitHub, Hacker News
Lack of real-world testing beyond iPhone 17 Pro
Seen on GitHub
Learning curve
intermediateProductive in ~A few hours
Hidden costs people mention
- • No hidden costs — but requires Apple developer account ($99/year) to deploy to devices
Viability Score
69/100
MonitorHow likely is coreai-model-zoo to still be operational in 12 months? Based on 4 signals — momentum (how recently it shipped), wrapper dependency, revenue model, and web presence.
momentum
55
funding runway
40
website health
90
wrapper dependency
100
Last calculated: July 2026
How we score →Key Features
- Pre-converted .aimodel files for 30+ models
- End-to-end conversion scripts with gotcha documentation
- Custom Metal kernels for GPU/ANE acceleration
- Swift runner (CoreAIKit) for one-line model loading
- iOS 27 and macOS 27 Core AI compatibility
- Verified on iPhone 17 Pro GPU/ANE
- MoE and dense model support
- MLA (GLM-4.7-Flash) support
- 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)
- Embedding and reranker models for RAG
- Object detection models (not explicitly listed but implied by 'detection')
About coreai-model-zoo
FreeAdvancedNo APIMobile · Desktop
CoreAI-Model-Zoo is a community-driven GitHub repository offering end-to-end converted Large Language Models (LLMs) optimized for Apple's Core AI framework on iOS 27 and macOS 27. It provides downloadable .aimodel files, conversion scripts, custom Metal kernels, and a Swift runner for on-device inference. The primary audience is iOS/macOS developers and AI engineers deploying local LLMs on Apple devices, with models verified on iPhone 17 Pro GPU/ANE. The zoo includes a diverse range of models: text LLMs (Qwen3.5, Qwen3.6, GLM-4.7-Flash, Gemma 4, Ornith-1.0, etc.), vision-language models (Qwen3-VL, Holo2-4B, MiniCPM-V 4.6), audio models (Whisper large-v3-turbo, Qwen3-ASR, Parakeet-TDT), text-to-speech (Kokoro-82M, VoxCPM), and even a diffusion LLM (LLaDA-8B) and a ternary VLA robot model (BitVLA). It also provides embedding and reranker models for RAG. Key features include pre-converted .aimodel files, conversion scripts with documented gotchas, custom Metal kernels for GPU/ANE acceleration, a Swift runner through CoreAIKit (SPM) that enables one-line model loading, and per-model integration cards with code snippets. All models are verified on-device on iPhone 17 Pro. What sets CoreAI-Model-Zoo apart is its focus on verified on-device performance and deep knowledge base covering Core AI pitfalls, MoE/dense model support, and official-QAT int4 quantization. Unlike generic model zoos, it provides concrete guidance on avoiding conversion issues and includes custom kernel optimizations for Apple Silicon. The project is fully open-source under permissive licenses.
Behind the Verdict
CoreAI-Model-Zoo fills a genuine gap: converting open-source models to Apple Core AI is notoriously brittle, and the curated collection saves hours of trial-and-error. The inclusion of custom Metal kernels and conversion gotchas shows real engineering effort. We'd reach for this whenever a team needs to run LLMs on iPhone or Mac without cloud dependencies. Where it falls short is platform lock-in — if your stack includes Android or Windows, this zoo is irrelevant. Also, the documentation assumes you already understand Core AI concepts; beginners may find the knowledge base dense. Compared to alternatives like MLX Community or ollama, CoreAI-Model-Zoo is more narrowly focused on iOS/macOS Core AI but offers deeper per-model optimization. It's not a general-purpose model hub. In practice, the one-liner Swift API (CoreAIKit) is a standout — downloading and running a chat model in three lines is genuinely convenient. The audio and vision model support broadens use cases beyond text-only apps. Bottom line: If you're shipping an on-device AI feature for Apple devices, this is your best starting point. If you need cross-platform or commercial support, look elsewhere.
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.
iOS developer building a local chatbot app
You want to integrate a private, offline LLM into your iOS app for customer support
Outcome: Download Qwen3.5-2B .aimodel, use the Swift runner, and integrate inference with 30 lines of code; no API key needed
ML researcher porting a model to Core AI
You converted a PyTorch model to Core AI but faced crashes
Outcome: Reference the conversion gotchas section, apply recommended Metal kernel patches, and run successfully on iPhone 17 Pro
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
Qwen3.5-0.8BQwen3.5-2BQwen3.6-35B-A3B (MoE)Qwen3.6-27BGLM-4.7-Flash (MoE+MLA)Gemma 4 E2B (int4)Gemma 4 E4B (int4)Gemma 4 12BGemma 4 31BLFM2.5-1.2B
as of 2026-07-01
Limitations
- The zoo currently supports only a limited set of models (Qwen, Gemma, GLM).
- All models require iOS 27 or macOS 27 and are tested only on iPhone 17 Pro.
- Conversion scripts may need specific environment setup.
- No cloud API is provided; all inference is on-device.
as of 2026-07-01
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Annual total
Free
Over 12 months
Effective monthly
Free
Billed monthly
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
Any developer with Apple hardware wanting to run local LLMs without cost
What this tier adds
Starting point: all models, scripts, and documentation are free under open-source licenses
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 completely free, open-source, and self-serve — you pay nothing, making it ideal for indie developers, students, and researchers who have Apple hardware. There are no paid tiers, unlike cloud-based LLM API services like OpenAI or Anthropic.
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 most developers: under 10 minutes to download a model and run the Swift runner on iPhone 17 Pro or Mac. Complex conversions may take a few hours to study gotchas and adjust scripts.
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.
Migrating in
- →From CoreML-Models: Migrate to CoreAI-Model-Zoo for newer model formats and better performance on iOS 27/macOS 27
- →From Hugging Face: Use provided conversion scripts to turn any HF model into .aimodel for on-device deployment
Migrating out
- ↗To Hugging Face: Convert .aimodel files back to PyTorch using provided reverse scripts (if available)
- ↗To llama.cpp: Re-export models from original sources for cross-platform GPU support
Resources & Guides
Frequently Asked Questions
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
Alternatives to coreai-model-zoo
View allUsed coreai-model-zoo? Help shape our editorial sentiment research.