Skip to content

On-device models

Models you can run on iPhone, iPad and Mac, from Hugging Face

Core ML and MLX models published on Hugging Face, plus the models Apple publishes there, sorted by downloads in the last 30 days. Core ML models run on iPhone, iPad and Mac through Apple's Core ML framework; MLX models are made for Apple's MLX library, mostly on Apple silicon Macs (MLX Swift also runs on iPhone and iPad). Each name links to the model's page on Hugging Face.

1,009 Core ML models, 1,693 MLX models and 143 from Apple, updated daily from Hugging Face's public API. Last updated 2026-11-02 (UTC).

New this month

The most liked models created in the last 30 days, across Core ML, MLX and Apple's own.

See all new models

Showing 77 of 77.

ModelTaskSizeDownloads (30 days)
aufklarer/Kokoro-82M-CoreMLCore MLfrom hexgrad/Kokoro-82MText to speech335 MB23.2K
FluidInference/kokoro-82m-coremlCore MLfrom hexgrad/Kokoro-82MText to speech5.6 GB16.1K
mattmireles/kokoro-coremlCore MLfrom hexgrad/Kokoro-82MText to speech3.4 GB6,909
aufklarer/Supertonic-3-CoreMLCore MLfrom Supertone/supertonic-3Text to speech—4,438
FluidInference/pocket-tts-coremlCore MLfrom kyutai/pocket-ttsText to speech—3,845
AILogDev/Irodori-TTS-v4.1-Small-MF-CoreMLCore MLfrom Aratako/Irodori-TTS-v4.1-Small-MFText to speech—2,671
FluidInference/supertonic-3-coremlCore MLfrom Supertone/supertonic-3Text to speech—2,483
iliasaz/chatterbox-nano-coremlCore MLfrom ResembleAI/chatterbox-nanoText to speech—1,979
seanll95/chatterbox-nano-coremlCore MLfrom ResembleAI/chatterbox-nanoText to speech—1,171
iliasaz/chatterbox-multilingual-coremlCore MLfrom ResembleAI/chatterboxText to speech—1,154
aufklarer/Supertonic-3-CoreML-FP16Core MLfrom Supertone/supertonic-3Text to speech—1,023
iliasaz/chatterbox-turbo-coremlCore MLfrom ResembleAI/chatterbox-turboText to speech—898
yepher/screen-cut-pro-tts-coremlCore MLfrom ResembleAI/chatterboxText to speech—719
iliasaz/ruaccent-coremlCore MLText to speech—464
mweinbach/Kokoro-82M-SwiftCore MLfrom hexgrad/Kokoro-82MText to speech—313
qmohits/kokoro-82m-coremlCore MLfrom hexgrad/Kokoro-82MText to speech—302
spybyscript/scyllasbandCore MLText to speech—300
qmohits/kokoro-82m-coreml-size-tiersCore MLfrom hexgrad/Kokoro-82MText to speech—261
kensora/kokoro-coremlCore MLText to speech—215
FluidInference/paradee-8m-coremlCore MLfrom sahilmahendrakar/Paradee-8M-v1.0Text to speech—210
anayak16/kokoro-coreml-int8Core MLfrom hexgrad/Kokoro-82MText to speech—179
arata-ae/negen-archiveCore MLfrom AILogDev/Irodori-TTS-v4.1-Small-MF-CoreMLText to speech—175
FluidInference/chatterbox-nano-coremlCore MLfrom ResembleAI/chatterbox-nanoText to speech—165
mosiah/phonon-31m-en-coremlCore MLfrom gradium/phonon-31m-enText to speech—161
itsmechico/chatterbox-multilingual-pt-br-coreml-aneCore MLfrom ResembleAI/Chatterbox-Multilingual-pt-brText to speech—143
itsmechico/moss-tts-nano-coreml-aneCore MLfrom OpenMOSS-Team/MOSS-Audio-Tokenizer-NanoText to speech—124
aufklarer/Magpie-TTS-Multilingual-357M-CoreML-8bitCore MLfrom nvidia/magpie_tts_multilingual_357mText to speech—111
itsmechico/qwen3-tts-0.6b-base-coreml-aneCore MLfrom Qwen/Qwen3-TTS-12Hz-0.6B-BaseText to speech—105
aufklarer/MOSS-TTS-Nano-100M-CoreMLCore MLfrom Allinsama/OpenMOSS-Team__MOSS-TTS-Nano-100MText to speech—104
robbiemu/paper2-kokoro-coreml-componentsCore MLfrom hexgrad/Kokoro-82MText to speech—89
s2e-host/s2e-kokoro-coremlCore MLfrom hexgrad/Kokoro-82MText to speech—86
FluidInference/chatterbox-multilingual-coremlCore MLfrom ResembleAI/chatterboxText to speech—82
FluidInference/moss-tts-nano-coremlCore MLfrom OpenMOSS-Team/MOSS-Audio-Tokenizer-NanoText to speech—77
aufklarer/Kokoro-82M-German-Martin-CoreMLCore MLfrom kikiri-tts/kikiri-german-martinText to speech—76
FluidInference/inflect-v2-coremlCore MLfrom owensong/Inflect-Micro-v2Text to speech—75
FluidInference/neutts-2e-coremlCore MLfrom neuphonic/neutts-2eText to speech—73
aufklarer/Kokoro-82M-CoreML-INT8Core MLfrom hexgrad/Kokoro-82MText to speech—72
seba/VoxCPM2ANE-PreviewCore MLfrom openbmb/VoxCPM2Text to speech—72
ronbakker/chatterbox-nano-int4-int8-coremlCore MLfrom ResembleAI/chatterbox-nanoText to speech—71
FluidInference/StyleTTS-2-coremlCore MLfrom yl4579/StyleTTS2-LibriTTSText to speech—70
aufklarer/Kokoro-82M-German-Victoria-CoreMLCore MLfrom kikiri-tts/kikiri-german-victoriaText to speech—70
wakirin/supertonic-3-coremlCore MLText to speech—69
aufklarer/Chatterbox-Flash-CoreMLCore MLfrom ResembleAI/chatterbox-flashText to speech—66
elphi-si/elphiCore MLfrom canberkkkkkk/ema-lightningText to speech—66
raratu/Onsei-iOS-ModelsCore MLText to speech—55
loudreader/loudr-1-turboCore MLText to speech—53
loudreader/loudr-1Core MLText to speech—51
agelwarg/supertonic-3-coremlCore MLfrom Supertone/supertonic-3Text to speech—50
FluidInference/luxtts-coremlCore MLfrom YatharthS/LuxTTSText to speech—47
AILogDev/sbv2-coreml-jvnv-f1-jpCore MLfrom litagin/style_bert_vits2_jvnvText to speech—46
FluidInference/qwen3-tts-coremlCore MLText to speech—46
okayuji/Qwen3-TTS-12Hz-1.7B-CustomVoice-CoreMLCore MLfrom Qwen/Qwen3-TTS-12Hz-1.7B-CustomVoiceText to speech—45
aufklarer/Pocket-TTS-100M-CoreML-INT8Core MLfrom kyutai/pocket-ttsText to speech—40
smdesai/kokoro-82m-coremlCore MLfrom hexgrad/Kokoro-82MText to speech—38
blueblaine/lector-voicesCore MLfrom hexgrad/Kokoro-82MText to speech—37
itsmechico/magpie-tts-multilingual-357m-coreml-aneCore MLfrom nvidia/magpie_tts_multilingual_357mText to speech—37
FluidInference/CosyVoice3-0.5B-coremlCore MLfrom FunAudioLLM/Fun-CosyVoice3-0.5B-2512Text to speech—36
Nooder/supertonic-2-coremlCore MLText to speech—36
itsmechico/phonon-40m-pt-coremlCore MLfrom gradium/phonon-40m-ptText to speech—35
aoiandroid/OpenVoiceV2-CoreML-mirrorCore MLText to speech—32
itsmechico/phonon-40m-es-coremlCore MLfrom gradium/phonon-40m-esText to speech—32
itsmechico/phonon-40m-fr-coremlCore MLfrom gradium/phonon-40m-frText to speech—32
itsmechico/phonon-40m-de-coremlCore MLfrom gradium/phonon-40m-deText to speech—31
itsmechico/phonon-40m-en-coremlCore MLfrom gradium/phonon-40m-enText to speech—31
FluidInference/magpie-tts-multilingual-357m-coremlCore MLfrom nvidia/magpie_tts_multilingual_357mText to speech—28
42futures/plapre-pico-coremlCore MLfrom syvai/plapre-picoText to speech—26
smdesai/Chatterbox-Multilingual-TTS-8bitCore MLfrom ResembleAI/chatterboxText to speech—25
seba/VoxCPM1.5-ANECore MLfrom openbmb/VoxCPM1.5Text to speech—23
erjigit17/Qwen3-TTS-0.6B-ANECore MLfrom Qwen/Qwen3-TTS-12Hz-0.6B-CustomVoiceText to speech—22
mlboydaisuke/Kokoro-82M-CoreAICore MLfrom hexgrad/Kokoro-82MText to speech—22
mlboydaisuke/VoxCPM-0.5B-CoreAICore MLfrom openbmb/VoxCPM-0.5BText to speech—21
ronbakker/chatterbox-nano-fp16-coremlCore MLfrom ResembleAI/chatterbox-nanoText to speech—21
aoiandroid/kokoro-82m-coremlCore MLfrom hexgrad/Kokoro-82MText to speech—20
seanll95/chatterbox-nano-long-utterance-en_gbCore MLfrom ResembleAI/chatterbox-nanoText to speech—20
aoiandroid/mirror-FluidInference-kokoro-82m-coremlCore MLText to speech—18
robbiemu/paper2-dia1-dac-coreml-decodersCore MLfrom descript/dac_44khzText to speech—18
itsmechico/omnivoice-coreml-aneCore MLText to speech—17

Browse by task

Pages for tasks with at least 5 models, across Core ML, MLX and Apple's own.

More on on-device models

How to run these

  • Core ML: add the .mlpackage to your Xcode project, or download a compiled model when the app first needs it, and load it with the Core ML framework. Some repos ship their own Swift package; the model card says which.
  • MLX: on a Mac with Apple silicon, the mlx-lm and mlx-audio Python packages load these by name; inside an app, MLX Swift runs them on Mac, iPhone and iPad. Mind the memory: in our test a 4-bit 0.5B model peaked at about 0.36 GB on a Mac; larger models need far more.
  • Apple's own: some are Core ML files ready for an app; others are research weights for PyTorch and need converting with coremltools first.

Reading the table

  • Downloads are Hugging Face's count for the last 30 days, not all time. Hugging Face counts requests for a model's main file, so one install can count more than once.
  • Size is the total of the repo's current files on Hugging Face, in the sizes it shows (1 GB = 1,000 MB). Some repos hold several variants of a model, so it can be larger than the one download you need; a dash means it has not been measured yet.
  • Trending is Hugging Face's own score for recent attention. "from" under a name is the model it was made from.
  • Parameters is the parameter count in the model files. A count marked * comes from the model's name (such as 27B) instead: the files gave none, or a much smaller one, as packed low-bit files can.
  • Quantization is read from the model's tags and name, not checked against the files.

Before you use one

The list comes from Hugging Face's public API and is updated daily. We don't test each model. Check the license and read the model card before you ship one in an app: some licenses limit commercial use, and "other" means the terms are on the model card. Models whose names or tags point at adult or safety-removed content are left out, and so are gated models that need a Hugging Face sign-in to download.