Skip to content
localmodel.run

Audio model · bark

BK Bark requirements

Text to speech (and sound effects) · 900M params · fp32 / fp16 · released Apr 2023. Light enough to run on CPU, no GPU required.

MIT Commercial use OK
Peak memory (fp32)
~5 GB
Runs on CPU
Yes
Parameters
900M
Type
Text to speech (and sound effects)

Run it

Runtime tools fp32

Bark runs in HF Transformers, suno-ai/bark or BetterTransformer at fp32. It runs CPU-only, and the smaller tiers are fast enough for real-time use on a laptop.

HF Transformerssuno-ai/barkBetterTransformer

Which devices can run Bark?

Apple Silicon Macs

RAM-only laptops

iPhone & iPad

No mainstream local runtime for Bark on iPhone & iPad yet.

Android

No mainstream local runtime for Bark on Android yet.

NVIDIA GPUs

AMD GPUs

FAQ

How much memory does Bark need?

At fp32 it consumes ~5 GB. It runs on CPU, so a GPU is optional.

Can Bark run on a phone or CPU?

Yes for CPU. It runs on Mac or laptop CPU, though no phone runtime is confirmed.

Can I use Bark commercially?

Yes. Bark is licensed MIT, which permits commercial use.

Notes

Suno's 3-stage generative audio model (~300M per stage, ~0.9B total) that can produce speech, music and sound effects. Peaks at ~5GB VRAM in fp32 via HF Transformers; drops to ~1GB with fp16 + CPU offload, and bark-small needs ~1.9GB. CPU-only works but needs ~8GB RAM and is slow. MIT, commercial OK. Sources: HF model card, the optimizing-bark guide, suno-ai repo.

Sources

Memory is a sourced peak-usage anchor at fp32, validated 2026-06-15. See methodology.