H3 Max

5-15 second cinematic video; 768p, 24fps, with audio, 6 aspect ratios.

modellane/h3-max · Text to video · MiniMax

MiniMax's cinematic video model. It produces 5-15 second, 24fps videos with audio from text. With 480p and 768p resolution options, six aspect ratios, and optional prompt expansion it fits everything from social media to ad clips. It can produce a 5-second clip in roughly 3 seconds — which speeds the jump from idea to result.

Vendor: MiniMax
Task: Text to video
Output: Video
Price: $0.06 · per second
Longest run: 15 seconds

Input parameters

Prompt prompt — required, at most 7000 characters Describe what you want produced as concretely as you can.
Duration duration — optional, default: 5, range: 5–15 sn, affects price A longer duration raises the price directly.
Resolution resolution — optional, default: 768p, options: 480p, 768p Higher resolution raises the per-second price.
Aspect ratio ratio — optional, default: 16:9, options: 21:9, 16:9, 4:3, 1:1, 3:4, 9:16 "adaptive" picks the ratio from the source.
Prompt expansion prompt_expansion — optional, advanced, default: false When on, the model enriches the prompt with its own rules.

Recommended use cases

Social media ad clips — Product or place promos in short, vertical, and square formats.
Cinematic scene tests — See a scene's light, camera, and pace before committing to production.
Fast draft production — A 5-second clip in 3 seconds; test an idea in motion right away.
Multi-format delivery — Produce 21:9 cinematic, 16:9 horizontal, and 9:16 vertical from the same prompt.

Where it is strong

Very fast — a 5-second clip is ready in about 3 seconds
Six aspect ratios, from 21:9 to 9:16
Synchronised stereo audio on every clip
24fps cinematic frame rate

Limits

No 2K resolution — 768p is the ceiling
15-second ceiling; not suitable for long scenes
No reference image, video, or audio support

Example prompt

A street food vendor in Istanbul at night, steam rising from the grill, hands working fast, neon signs reflecting in wet cobblestones, shallow depth of field, cinematic 24fps

Example request

POST /v1/queue/modellane/h3-max

Playground

The identifier, price, parameters, and example requests on this page can be read without signing in. Running the model from the panel sends a real generation request and charges the project balance; the playground therefore requires an account.