Phosor

MiniMax H3 Standard & Turbo

AI video generator. Start with what you have.

Create AI video with MiniMax H3 Standard or Turbo. Start with words, a still image or visual references, then create in the Phosor playground or build with the API.

Try MiniMax H3 Find your starting point

Browse without an account. Sign in to generate. Pay per use.

Choose how to make your video

Start with the input you need to control. A reference image and a starting frame serve different purposes.

01

Text to video

Describe a scene from scratch when you do not have a starting image. Good for exploring a setting, an action or a camera movement before committing to a visual direction.

What you need
A written prompt describing the subject, setting and motion.
What it guides
Your words guide the appearance and action. For a specific existing subject, consider an image or references instead.

H3 API input fields

prompt

Begin with one subject, one action and one camera instruction. Build a more complex sequence after reviewing a simple shot.

MiniMax H3 Text to video MiniMax H3 Text to video Turbo
02

Image to video

Animate an image you already have: a product photograph, illustration or planned first frame. Describe how the subject and camera should move from that starting composition.

What you need
A prompt and a starting image. An ending image is optional.
What it guides
The image defines the opening frame. An optional ending image guides where the shot should finish.

H3 API input fields

promptimage_urlend_image_url

Use image_url for the first frame and end_image_url for the optional last frame. Choose endpoints that can connect through a plausible motion.

MiniMax H3 Image to video MiniMax H3 Image to video Turbo
03

Reference to video

Use reference images or clips to guide a subject, location or movement. Optional audio references can guide sound. Tell H3 which role each uploaded reference should play.

What you need
A prompt and at least one reference image or video. Audio references are optional; audio alone is not enough.
What it guides
References guide the scene without defining an exact opening frame. Use image to video when you need a specific first frame.

H3 API input fields

promptreference_image_urlsreference_video_urlsreference_audio_urls

Address references with <Picture n>, <Video n> and <Audio n>. Match each token to the order shown in the playground. Check current upload and reference limits there.

MiniMax H3 Reference to video MiniMax H3 Reference to video Turbo

Written prompt examples

Starting points to adapt to your own shot. These are prompt examples, not generated video samples.

Text to video

A lone hiker walks along a snowy mountain ridge at sunrise. A slow tracking shot follows from behind. Fine snow blows across the path; warm light catches the peaks.
Check the required inputs

Image to video

Starting from the supplied image, the fabric moves gently in a light breeze. The camera slowly pushes toward the subject. Maintain soft daylight and the original setting.
Check the required inputs

Reference to video

Use the person from <Picture 1> in the garden shown in <Picture 2>. The person walks slowly toward the fountain. A steady medium shot follows; leaves move in a light breeze.
Check the required inputs

Review the shot before building on it

Check your result against the intention of the shot, rather than judging only a single attractive frame.

  • Subject: are the person, product and important visual details recognizable throughout the clip?
  • Motion: do the action, camera movement and first-to-last-frame transition read clearly?
  • Details: inspect faces, hands, text, edges and continuity across several frames, and listen to any audio.

If the shot needs work, change one part of the prompt or input at a time so you can compare the difference.

Compare H3 Standard and Turbo

From playground to API

Choose a mode in the playground, then use its exact public model ID in your integration. Generation is asynchronous: submit a request, check its status and retrieve the completed result.

H3 Standard API guide · H3 Turbo API guide

  1. Create an API key in your account.
  2. Submit the model ID, prompt and required input fields.
  3. Check request status → Get the completed result
API documentation

Check the full cost of the shot

H3 output is billed by video duration. Rates depend on resolution and variant; reference inputs may add charges. Review the current output and reference prices before submitting. Compare Standard and Turbo with your own prompts and assets.

View current pricing

Before you start

Can I generate AI video without code?

Yes. Choose an H3 model and sign in to create in the playground. Generation is billed per use.

Open the H3 playground
How do I turn a photo into a video?

Use image to video with a starting image and a motion prompt. You can add an optional ending image to guide the final frame.

Try image to video
Should I choose H3 Standard or Turbo?

Both offer text, image and reference to video. Compare current resolution and reference-input rates, then judge each with your own prompts and assets.

Compare current pricing
How do I use the AI video generation API?

Create an API key, submit the exact H3 model ID and required inputs, then check status and retrieve the result using the returned request_id.

Read the H3 API guide