milveyVideo and GPUs
PricingModels
DocsJoin the list

Product

CreateWorkflowsDeployGPUsPricing

Discover

ModelsLoRAsCommunityExploreChangelogMedia sourcesTermsPrivacyRefunds

Developers

DocsAPI referenceQuickstart

Account

Join the list

Join the list

I’ll email you when signups open.

Unsubscribe any time. See the privacy policy.

ChangelogDocsAPI referencePricing
© 2026 Milvey · In early buildTermsPrivacyRefundsMedia sources
milveyvideo and gpusmilveyvideo and gpusmilveyvideo and gpusmilveyvideo and gpus

hkchengrex (Sony AI / UIUC)

MMAudio

Repo examples are NC: not downloaded. 2 credits per 5 s, about 10-30 s per run. Sample price.

Join the list Compare all models
MMAudio spec
StatusBeta
TypeAudio
Versionlarge_44k_v2
Durations1-30 s (fal)
AudioNative audio track
LicenceCode MIT; weights CC-BY-NC-4.0
Where it runsMilvey GPUs. The run queues if the pool is full.
Typical time10-30 s
Price2 cr per 5 s

Prices and times are samples.

A 1,000 cr pack buys

500 clipssample

At 2 cr per 5 s. Credit prices

  • Video-to-audio foley synced to video
  • Text-to-audio
  • 44.1 kHz

MMAudio runs on Milvey GPUs.

Open models run on Milvey GPUs, and the price follows that GPU time. If the pool is full, the run waits in a queue until capacity returns. It never moves to a paid API. If the run can’t be delivered, the hold is released and you pay 0 cr.

How failed runs are charged
  1. Milvey GPU pool10-30 s

MMAudio makes sound, so there is nothing to show here

Sources and licences

Samples on this site are the makers’ own published outputs, each credited with its source. hkchengrex (Sony AI / UIUC) has none listed here so far. Outputs made on Milvey will be added as the beta runs.

Licence: Code MIT; weights CC-BY-NC-4.0

Not available for paid runs

Not for paid runs. Code MIT; weights CC-BY-NC-4.0.

Known unknowns

Nothing flagged for this model.

Sources for specs and licence

  • huggingface.co/hkchengrex/MMAudio
  • fal.ai/models/fal-ai/mmaudio-v2

Other audio models

ElevenLabs

ElevenLabs v4 TTS + Eleven Music20 cr

Voice cloning terms not reviewed.

Seconds (TTS); 30-90 s (music) · 2 routes · audio

per 1k characters

Stability AI

Stable Audio 2.550 cr

Music and SFX. Clips up to 190 s.

Seconds to tens of seconds · 1 route · audio

per generation

Try MMAudio with 150 free credits.

Join the list. I’ll email you when signups open. No card needed.

Email when signups open · Unsubscribe any time

MMAudio · Milvey