
MiniMax H3 Max Uncensored
MiniMax H3 Max Uncensored is MiniMax's video model, hosted on Vellria with nothing to install: text to video and image to video, 480p to 4K, clips of 5–15s. Run it in Studio or over the REST API, from 15 credits / s (100 credits = $1.00).
vellria-minimax-h3-maxMiniMax H3 Max is developed by MiniMax. Vellria provides access to the model; it does not own it. All trademarks belong to their respective owners.
Output
No generation yet
Cover art
from 15 credits / s
Your request will cost 125 credits.
About MiniMax H3 Max
Also known as: H3 Max · MiniMax H3 Max video
MiniMax H3 Max is the top tier of MiniMax's H3 video line — the company behind Hailuo AI. Where MiniMax H3 Video is the scratchpad, Max is for the take you mean to keep. The two share the same resolutions and frame shapes, so what you settle on the standard tier comes back here at the same size and only the price changes.
That shared geometry is the reason to use them as a pair. Work out the shot on the standard tier: the framing, the camera move, the length, the moment the action peaks. When a draft is right, run the identical prompt here. You are not rediscovering the scene at a new tier; you are rendering a decided shot with the better model.
Every clip comes with sound, so write the action to be heard as well as seen — fabric, breath, footsteps, the room. For the final pass, the most dependable input is usually a still you have already approved: the clip opens on that exact frame and inherits its proportions, so the look you signed off on is the look you get and nothing needs reframing after rendering. A text-only prompt still works when no picture exists yet, and then you set the proportions yourself.
If you are deciding between H3 Max and Seedance 2.5 for a final clip, run the same short prompt on both: they differ in how they pace movement and render sound, and the better match depends on the scene. For long unbroken takes, Wan 3.0 Pro is the other finishing model.
MiniMax H3 Max in action
MiniMax H3 Max outputs with the exact prompts behind them.
Image to video
The singer sways and sings into the microphone, the spotlight shimmering on her sequins, smoke drifting, slow camera push-in
Image to video
He pushes off the doorframe and steps toward the camera, light sweeping across his torso, slow dramatic motion
Text to video
A masquerade ball in a grand baroque hall, couples in their thirties in gowns and tuxedos waltzing under crystal chandeliers, the camera glides between the dancers
When to pick MiniMax H3 Max
Pick it when
- You already settled framing, camera move and length on MiniMax H3 Video and want the identical prompt rendered as the final take.
- The clip will be published as is and every beat should carry sound, from fabric and breath to footsteps and room tone.
- You have an approved still and want the final clip to open on that exact frame at a matching resolution.
Pick another model when
- You are still exploring the idea, camera or pacing; draft on the cheaper MiniMax H3 Video first.
- The finished clip needs one long unbroken take rather than shorter beats; Wan 3.0 Pro is the finishing model for that.
- Sound design should carry the moment; run the same short prompt on Seedance 2.5 and compare.
Prompt or reference
Write the scene from scratch when the shot has no picture behind it yet, and choose the frame shape yourself. Start from a still when you already have the look: it opens the clip and fixes its shape, so crop before you upload rather than after the render.
How to use MiniMax H3 Max on Vellria
- Prompt: up to 7,000 characters in the panel above.
- Cost: from 15 credits / s — see pricing.
Example prompts
Final take: an adult woman steps out of a claw-foot bath, wraps herself in a towel and walks to a fogged mirror, candlelight flicker, water dripping, slow tracking shot.
From the uploaded still: the adult couple on the balcony kiss as fireworks burst behind them, crowd noise far below, camera cranes up.
An adult man in an open shirt plays a slow piano piece in an empty bar at closing time, the bartender's glass clinks, warm amber light, gentle push-in.
Adult dancer in a sequined gown finishes a spin and drops into a bow on a smoky stage, applause swells, spotlight narrows.
Tips
- Never explore on Max; finalize the prompt on H3 Video at the same resolution first.
- Keep the wording identical between tiers so the result changes only in finish, not in content.
- Describe at least one sound per beat; every clip includes audio.
- For image-to-video, upload the exact still you plan to publish as the opening frame.
Specifications
| Resolution: | 480p768p1440p4K |
|---|---|
| Duration: | 5–15s |
| Aspect ratio (Text to video): | 21:916:94:31:13:49:16 |
| Reference images: | 1 reference image |
See also: uncensored AI video generator · NSFW video generation · NSFW image to video.
Pricing
| Resolution | Credits |
|---|---|
| 480p | 15 credits / s |
| 768p | 25 credits / s |
| 1440p | 40 credits / s |
| 4K | 60 credits / s |
MiniMax H3 Max vs MiniMax H3 Video and MiniMax H3 Image
MiniMax H3 Max is the finishing tier of the H3 line. It shares every resolution step and frame shape with H3 Video, so the usual path is to draft there and render the keeper here.
| Model | Price | Max resolution | Duration | Reference images |
|---|---|---|---|---|
| MiniMax H3 Max Uncensored This page | from 15 credits / s | 4K | 5–15s | Up to 1 |
| MiniMax H3 Video Uncensored | from 6 credits / s | 4K | 4–15s | Up to 1 |
Prices and limits come from the live model catalog.
MiniMax H3 Max or MiniMax H3 Video?
- Price: from 6 credits / s (MiniMax H3 Max: from 15 credits / s)
- Duration: 4–15s (MiniMax H3 Max: 5–15s)
Draft on H3 Video while framing, camera and pacing are still open; it costs less per second at every resolution. Bring the prompt back here unchanged once a take works.
MiniMax H3 Max or MiniMax H3 Image?
Image · 8 credits
H3 Image makes the still that opens the clip. A frame designed there keeps the same family look when Max animates it.
MiniMax H3 Max API: price and a first request
MiniMax H3 Max costs from 15 credits / s over the API, the same rate as in Studio. Model id vellria-minimax-h3-max on POST /v1/generate/video.
curl https://vellria.com/v1/generate/video \
-H "Authorization: Bearer vll_..." \
-H "Content-Type: application/json" \
-d '{
"model": "vellria-minimax-h3-max",
"prompt": "…",
"duration": 5,
"resolution": "768P",
"ratio": "16:9"
}'Parameters, polling and uploads: API documentation · NSFW video API.
FAQ
Does MiniMax H3 Max allow NSFW?
On Vellria, yes, for fictional adult content. Nudity and explicit scenes between adults are allowed. Minors are refused automatically and real people may not be used without consent.
Is MiniMax H3 Max free?
Free to start: a new account created with Google gets 20 free credits, valid for 7 days (about 10 images on Z-Image Turbo Uncensored), and every signed-in account gets free daily uses of the NSFW prompt generator and the image to prompt tool. The gift can go toward MiniMax H3 Max at from 15 credits / s; beyond that every run is paid from prepaid credits at the rate this page shows before you start, and a run that fails at the provider is refunded automatically. Is there a free uncensored AI generator? covers the cheapest way to a first result.
Should I use H3 Max or H3 Video?
Explore on H3 Video, which shares resolutions and shapes with Max, then render the decided shot on Max.
Does H3 Max make sound?
Yes, every clip includes audio generated from your description.
Can I call H3 Max from the API?
Yes. The API section below shows the request for this model; the docs cover the rest.






