Now on VeniceVideoAnonymous

MiniMax H3 Hyper

The budget tier of MiniMax's H3 video family — 15-second 768P clips with native stereo audio, from $0.10 per clip on Venice.

For agents
curl https://api.venice.ai/api/v1/video/queue \
  -H "Authorization: Bearer $VENICE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "minimax-h3-hyper-text-to-video",
    "prompt": "Aerial drone shot over a misty mountain valley at golden hour",
    "duration": "5s",
    "resolution": "768P",
    "aspect_ratio": "16:9"
  }'

# Use the returned queue_id with https://api.venice.ai/api/v1/video/retrieve.
# Call /video/complete after downloading if needed.
Model IDminimax-h3-hyper-text-to-video
Maker
MiniMax
Modality
Video + audio
Max duration
15 seconds
Max resolution
768P

Overview

What is MiniMax H3 Hyper

MiniMax H3 Hyper is the budget tier of MiniMax's H3 omni-modal video family, live on Venice since October 2026. It generates 4–15 second clips at 768P with native 32 kHz stereo audio — dialogue, sound effects and music rendered in the same pass — in six aspect ratios, from text, a still image, or reference images.

Using it anonymously on Venice

On Venice, MiniMax H3 Hyper runs under the anonymized privacy tier: your prompts are not stored, profiled, or used for training, and there is no personal generation history tied to your clips. The model is also uncensored, so creative briefs that mainstream hosted video services quietly reject stay on the table. You pay per clip instead of a subscription, so a single 4-second 768P render costs $0.10.

AnonymousNo prompt training

Specifications

Datasheet

Maker
MiniMax
Open weights
No
License
Proprietary
Clip lengths
4s – 15s
Resolutions
768P
Mode
text-to-video
Audio
Yes
Prompt limit
7,000 chars
Released
July 31, 2026
Architecture
33B dense single-stream H3-Omni-Transformer with a Qwen3-VL-32B text encoder and H3-VAE for joint audio-video latents
Frame rate
24 FPS
Aspect ratios
16:9, 21:9, 4:3, 1:1, 3:4, 9:16
Privacy on Venice
Anonymized — prompts not stored
Available on Venice since
Oct 2026

Assessment

Strengths and limitations

Strengths
  • Joint audio-video generation: native 32 kHz stereo sound — speech, sound effects, and background music — rendered in the same inference pass as the frames, so dialogue stays lip-synced.
  • Three ways in on Venice: pure text prompts, animating a still image, or locking a subject's identity with reference images across shots.
  • Strong complex-instruction following and in-video text and brand rendering, per MiniMax's own launch materials — useful for ads and product content.
  • Clips run up to 15 seconds at 24 FPS with six aspect ratios from cinematic 21:9 to vertical 9:16.
  • The cheapest entry into the H3 family on Venice: 768P clips start at $0.10 — and prompts are never stored.
  • Uncensored on Venice: no content-policy gatekeeping layer between you and the model.
Limitations
  • Capped at 768P on Venice: the full H3 system can reach 2K through a separate regeneration module, but that stage is not offered here, and most Venice rivals output 1080p.
  • No open weights: you cannot self-host or fine-tune it, so Venice access is per-clip API usage only.
  • 15-second maximum per take; if you need a single 16–30 second shot, other models fit better.
  • Third-party side-by-sides found rivals like Seedance 2.5 follow certain directional action prompts more reliably — H3 Hyper trades some controllability for price.

Use cases

What it is good for

  1. 01Short ad creatives with a scripted line spoken by an on-screen character, delivered with audio in one render.
  2. 02Product and e-commerce clips animated from a single still image.
  3. 03Character-consistent social series built with the reference-to-video variant.
  4. 04Vertical 9:16 sound-on social content where a $0.10 draft beats a subscription minimum.

Prompting

Getting better results

Set the duration explicitly — any whole second from 4 to 15 — and draft at 4–6 seconds before committing to a 15-second take.

Put spoken lines in quotes and describe the voice; audio is generated jointly, so the model renders dialogue, effects, and music from the same prompt.

Describe the soundscape on purpose ("rain on pavement, distant traffic, no music") or the model will fill the track for you.

Pick the aspect ratio to match the surface: 9:16 for Reels/TikTok, 1:1 for feed posts, 21:9 for cinematic b-roll.

For recurring characters, switch to the reference-to-video variant and supply clean, well-lit subject images rather than re-describing the character in text.

Use the full 7,000-character budget for complex shots — camera moves, lighting, and timing all benefit from explicit description.

Samples

Sample outputs

Generated on Venice with our standard prompt suite — the same prompts we run through every model of this type, so you can judge it like-for-like.

Cinematic landscape

Slow aerial drone shot gliding over a misty mountain valley at golden hour, sunlight piercing clouds onto a winding river, ancient pine forests on either side, ultra-smooth motion, professional color grading, atmospheric haze, 4K

Seamless loop

Calm ocean waves rolling onto a black-sand beach at sunrise, golden light on wet sand, foam dissolving into the shore, a single silhouetted figure at the waterline, smooth continuous forward push, serene cinematic atmosphere

Urban cinematic

A slow tracking shot through a rain-soaked Tokyo street at night, neon reflecting in puddles, steam rising from a food cart, a person with a translucent umbrella, shallow depth of field, teal-and-orange grade, smooth steady camera

Compare every video model on these prompts →

Alternatives

How it compares

ModelBest forMax durationMax resolutionNative audio
MiniMax H3 HyperCheapest audio-visual clips15s768PYes
MiniMax H3 MaxHighest-fidelity H3 output15s1080PYes
Grok Imagine 1.5Fast stylized clips15s1080pYes
Wan 2.7 EnhancedOpen-weights alternative15s1080pYes
Vidu Q3Longest single take16s1080pYes

MiniMax H3 Hyper is the right pick when you need sound-on, subject-consistent clips up to 15 seconds at the lowest per-clip price on Venice — step up to H3 Max only when 1080P outweighs cost.

API

Call it from your code

Venice exposes this model through the REST API. Queue a generation with the model id.

curl https://api.venice.ai/api/v1/video/queue \
  -H "Authorization: Bearer $VENICE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "minimax-h3-hyper-text-to-video",
    "prompt": "Aerial drone shot over a misty mountain valley at golden hour",
    "duration": "5s",
    "resolution": "768P",
    "aspect_ratio": "16:9"
  }'

# Use the returned queue_id with https://api.venice.ai/api/v1/video/retrieve.
# Call /video/complete after downloading if needed.

Pricing

What it costs on Venice

Pay per clip on Venice — price scales with resolution and duration (4s–15s), from $0.10.

768P · 4s
$0.10
Per clip

New Venice accounts include a free daily allowance and 500 welcome credits — no credit card required.

FAQ

Frequently asked questions

MiniMax H3 Hyper is the budget tier of MiniMax's H3 omni-modal video family, available on Venice since October 2026. It generates 4–15 second clips at 768P and 24 FPS with native 32 kHz stereo audio, and ships in three modes: text-to-video, image-to-video, and reference-to-video.

On Venice you pay per clip, with the price scaling by resolution and duration. A 768P 4-second clip starts at $0.10, and longer takes up to 15 seconds cost more. There is no subscription requirement — credits are consumed only when you render.

MiniMax H3 Hyper does not have open weights, so it cannot be self-hosted or fine-tuned. On Venice you pay per generated clip, and new accounts include free daily credits you can use to try it before paying.

On Venice, MiniMax H3 Hyper outputs 768P. The full H3 system from MiniMax can reach 2K through a separate regeneration module, but that stage is not part of the Venice offering — if you need 1080p or higher, MiniMax H3 Max or Grok Imagine 1.5 are the closer options on the platform.

Yes. The variant with model id minimax-h3-hyper-image-to-video animates a still image into motion, with the same 4–15 second clip lengths, 768P resolution, and native stereo audio as the text-to-video variant. There is also a reference-to-video variant (minimax-h3-hyper-reference-to-video) that keeps a subject consistent using reference images.

Both generate up to 15 seconds with native audio; H3 Max outputs 1080P while H3 Hyper outputs 768P at a lower per-clip price. Use H3 Hyper for drafts, social content, and high-volume rendering; use H3 Max when the final deliverable needs the extra resolution.

Between 4 and 15 seconds, selectable in one-second increments, at 24 FPS. Longer continuous shots require stitching multiple clips or a different model — Vidu Q3 on Venice adds a 16-second option.

Yes. Frames and 32 kHz stereo audio are generated jointly in a single pass, covering speech, sound effects, and background music. Writing dialogue in quotes in the prompt is the most reliable way to get it spoken on-screen.

Venice runs MiniMax H3 Hyper under its anonymized privacy tier: prompts are not stored, profiled, or used for training, and generations carry no personal account history. You can use it in the Venice app or through the Venice API without building a generation profile.

Use MiniMax H3 Hyper anonymously

Venice does not store your prompts. Chat history stays in your browser.