OmniHuman 1.5: what it does, what it costs

Audio-driven lip sync: supply a face and a script, and clip length follows the voice, up to 30 seconds.

OmniHuman 1.5 is one of the models you can run inside UGC Vids AI. It is made by ByteDance, renders clips up to 30 seconds, and we price it at 720p and 1080p. A render costs 490 to 980 credits depending on the length and resolution you pick, all from the same credit balance as every other model in the catalogue. Everything on this page is generated from the pricing table the app charges against.

OmniHuman 1.5 specs

Vendor: ByteDance
Max clip length: 30 seconds
Resolutions priced: 720p and 1080p
Clip lengths priced: 15s, 30s
Audio: Runs on a voice track we generate
Reference image: Required
Cheapest render: 490 credits (15s, 720p)
Most expensive render: 980 credits (30s, 1080p)
Free 3-day trial: Paid plans only
Catalogue name: OmniHuman 1.5

OmniHuman 1.5 credit costs, in full

Every length and resolution combination we price for OmniHuman 1.5. Credits per second is included so you can compare this model against one whose clip lengths do not line up with it.

Clip length	Resolution	Credits	Credits per second
15s	720p	490	32.7
15s	1080p	490	32.7
30s	720p	980	32.7
30s	1080p	980	32.7

The two resolution rows for OmniHuman 1.5 cost the same credits at every clip length because the request we send this model carries no resolution setting at all. The tier is a label on our side, so there is nothing to gain or lose by choosing one over the other.

Credits come from your monthly plan balance and are shared across every model. Starter is $49 a month with 5,000 credits, Growth is $99 with 12,000, and Agency is $199 with 25,000. See the pricing page for annual rates and what happens when a balance runs low.

What OmniHuman 1.5 is

OmniHuman 1.5 is ByteDance's audio-driven talking-head model. It has no duration input: you give it a still image and a voice track, and the output runs as long as the audio does. We generate that voice track from your script before dispatch, so from your side it is still image plus script in, lip-synced video out. This is the one model where we pin an exact upstream version rather than following the vendor's rolling default, so the build cannot change under you without a deploy.

Where OmniHuman 1.5 fits, and where it does not

Good for

Longer talking-head reads, up to 30 seconds of speech
Scripts where lip sync matters more than scene motion
Reusing one avatar photo across many different scripts

Limits

No duration control: length is whatever the voice track runs to, and billing follows the script's spoken length
Needs a face image, so there is no prompt-only path
Both priced tiers, 720p and 1080p, cost the same and the model receives no resolution setting at all, so the tier label does not change the render

Choose OmniHuman 1.5 when the deliverable is a person delivering a script and the read is longer than a scene model's ceiling.

How OmniHuman 1.5 compares

Each comparison puts both credit tables on one page, so you can see which model is cheaper for the exact render you have in mind rather than in the abstract.

OmniHuman 1.5 vs Happy Horse 1.1

Max clip 30s vs 15s. Cheapest render 490 vs 140 credits. Top resolution 1080p vs 1080p.

Back to the full model table or browse every model comparison.

How to use OmniHuman 1.5 here

Start an account. The first 3 days are free. OmniHuman 1.5 itself runs on any paid plan rather than on the trial. One credit balance covers every model you can run, so signing up does not tie you to one.
Open the create page and pick OmniHuman 1.5 from the model list.
Add your reference image, the still this model works from, then write the prompt or script. Your text reaches the model as you wrote it.
OmniHuman 1.5 takes no duration: the render runs for as long as the voice track does, and the rows above are the priced anchors the charge is worked out from. The credit cost is shown before you generate, so nothing is charged that you have not seen first.
Generate, then download the mp4. Renders land in your library and stay there.

OmniHuman 1.5 is driven by a voice track rather than a duration setting. We generate that track from your script before the render starts, so from your side it is an image plus a script.

OmniHuman 1.5 on the free 3-day trial

OmniHuman 1.5 is not available during the free 3-day trial. It runs on any paid plan, and the trial covers the rest of the catalogue in the meantime.

What a trial can run: Grok Video, Happy Horse 1.1, Kling 2.6, Pruna Avatar, Seedance 1.5 and Veo 3.1 Fast. Their pages carry the same spec and credit tables, so you can test the workflow end to end for free and switch to OmniHuman 1.5 on a paid plan without changing anything else. See the full model index.

OmniHuman 1.5: frequently asked questions

How much does OmniHuman 1.5 cost per render?

The priced anchors run from 490 to 980 credits at 15s and 30s, and the two resolution labels cost the same because the model takes no resolution setting. In practice the charge follows the script's spoken length, and the estimate is shown before you generate.

How long can an OmniHuman video be?

Up to 30 seconds, which is longer than any model where you pick the length will render in a single call, since those cap at 15 seconds. Length is set by the voice track rather than by a duration setting.

Does OmniHuman need an audio file?

The model is audio-driven, but you do not have to supply the audio yourself. We generate a voice track from your script and pass it to the model in the same request.

OmniHuman 1.5 or Happy Horse 1.1?

OmniHuman runs to 30 seconds and drives the mouth from a voice track. Happy Horse 1.1 has a shorter ceiling at 15 seconds but generates picture and audio jointly in one pass and accepts up to 9 reference images. For a long read, OmniHuman. For a short clip where you want reference control, Happy Horse.

Can I use OmniHuman 1.5 on the free trial?

OmniHuman 1.5 is not available during the free 3-day trial. It runs on any paid plan, and the trial covers the rest of the catalogue in the meantime. The models a trial can run are Grok Video, Happy Horse 1.1, Kling 2.6, Pruna Avatar, Seedance 1.5 and Veo 3.1 Fast.

Start free for 3 days, and run OmniHuman 1.5 on any paid plan

Start free for 3 days

Free for 3 days. Cancel anytime.