Model reference · open weights

speakerkit-coreml

Not for GPU servers Audio argmaxinc Speech→text 1 build Licence not stated 37k dl/mo

speakerkit-coreml is an open-weight audio or speech model from argmaxinc. Built for Apple devices (Core ML). It does not run on NVIDIA GPUs.

All audio & music

SpeakerKit Core ML is an automatic speech recognition model developed by argmaxinc.
It is designed for speech recognition tasks and is available as the Swift framework SpeakerKit, which uses an MIT license, while the underlying models are licensed under CC BY 4.

Summary of the argmaxinc/speakerkit-coreml model card, 2026-10-01

What it is

What it runs on

Memory and cards for speakerkit-coreml (BF16)

32 MBweights, file size

1.6 GBruntime overhead

From the model card

Read the model card

SpeakerKit Core ML

Read the blog

Read the Interspeech 2025 paper for architecture and benchmarks.

Quoted from the model card on Hugging Face. The full card is behind the Hugging Face link above.

How it works

Running it

Built for Apple devices (Core ML). It does not run on NVIDIA GPUs.

Details

Type	Audio & music
Task	Speech→text
Runs with	whisperkit
Popularity	37k downloads / month
Weights	32 MB, speakerkit-coreml (BF16, file size)