Source-backed guide
H3 Max Text to Video
A route-level guide for turning a prompt into a short H3 Max video through fal-compatible request fields.
This guide covers h3 max text to video.
Official fal model source: fal H3 Max Text to Video Max model page
Official fal API source: fal H3 Max text-to-video API reference
Official fal workflow source: fal guide: how to use MiniMax H3 Max
Official fal billing source: fal model pricing documentation
Read first
Route and request shape
H3 Max text to video on fal uses the route minimax/h3-max/text-to-video. The public API reference lists a prompt, duration, resolution, seed, safety checker, sync mode, prompt expansion mode, and aspect ratio fields, with output including a video object, expanded prompt, and timings. Source: Official fal API source.
The most reliable implementation pattern is queue-first. Submit the request from a server route, keep the provider credential server-side, poll status, then request the result URL when the job completes. h3maxapi.com's workbench mirrors that shape with a site-issued key in the browser and provider access handled by the server. Source: Official fal API source.
Official media
Official text-to-video sample
The embedded clip is fal-hosted media from the H3 Max Text to Video Max model page.
H3 Max text-to-video sample
Official fal sample output from the H3 Max Text to Video Max model page.
Source table
Core fields
These fields are the first things to expose in an application UI or server request builder.
| Field | Typical choice | Why it matters | Source |
|---|---|---|---|
| prompt | A specific scene, camera, subject, action, and style. | It is the main control surface for text-to-video. | Official fal API source |
| duration | Start with 5 seconds, then test longer clips. | Short clips control cost and reveal prompt quality quickly. | Official fal API source |
| resolution | 480P for drafts, 768P for closer review. | The selected resolution changes per-second price. | Official fal model source |
| aspect_ratio | 16:9, 9:16, 1:1, 4:3, 3:4, or 21:9. | Pick this before writing camera direction so composition matches delivery. | Official fal API source |
| sync_mode | Use the queue-first default for longer jobs. | Async status and result handling keeps browser requests from waiting on generation. | Official fal API source |
| prompt_expansion_mode | balanced, precise, none, or quality depending on the API option. | Expansion can improve prompt detail but may add processing time. | Official fal API source |
Practical reading
Prompt workflow
Good H3 Max text-to-video prompting is an editing loop, not one long paragraph thrown at the API.
One shot per request
Keep each request focused on one shot. Name the subject, environment, camera motion, and the action that should happen during the clip.
Official fal workflow source: fal guide: how to use MiniMax H3 Max
Draft at 480P
Use 480P and short duration for exploration, then move accepted prompts to 768P when detail matters.
Official fal workflow source: fal guide: how to use MiniMax H3 MaxOfficial fal model source: fal H3 Max Text to Video Max model page
Store prompt and result together
Save the prompt, expanded prompt, timings, model route, resolution, and result URL together so winning clips are reproducible.
Official fal API source: fal H3 Max text-to-video API reference
FAQ
H3 Max Text to Video questions
What is the H3 Max text-to-video route?
fal lists the route as minimax/h3-max/text-to-video.
Which aspect ratios are available?
The fal API reference lists common ratios including 16:9, 9:16, 1:1, 4:3, 3:4, and 21:9.
Should I call the route directly from browser code?
No. Keep provider credentials server-side and use a site-issued key or backend route for browser interactions.
Source map
Where this page gets its facts
Every hard claim on this page is tied to official fal material or an independent benchmark source. Source basis checked September 4, 2026.
fal H3 Max Text to Video Max model page
Used for the H3 Max text-to-video route, launch-rate pricing, sample output video, and public model-page positioning.
fal H3 Max text-to-video API reference
Used for request fields including prompt, duration, resolution, aspect ratio, prompt expansion, safety checker, queue submission, and result payload shape.
fal guide: how to use MiniMax H3 Max
Used for the practical workflow of testing 5 to 15 second clips, using 480P drafts, moving to 768P, and separating text, image, and reference modes.
fal model pricing documentation
Used for the rule that model billing depends on generated output, that failed server jobs and queue waiting are not charged, and that live model pages should be checked before paid use.