No video yet
Add a prompt to generate your first video.
MiniMax H3 generates AI video from a detailed text prompt, a single start image, or a pair of first and last frames. Describe the subject, setting, camera movement, lighting, and pacing, then choose a 4–15 second duration, aspect ratio, and either 768P or 2K output. It is a practical video workflow for campaign concepts, product storytelling, controlled frame-to-frame motion, and cinematic social content.
Prototype product reveals, campaign atmosphere, and landing-page hero motion before committing to a production shoot or a higher-cost video workflow.
Write a clear action arc for a character, then generate a short, shareable sequence with the framing and pacing suited to vertical or feed-native creative.
Turn a campaign still into a concise editorial movement study, with direction for subject motion, lens behavior, fabric, light, and pacing.
Specify composition, action, camera behavior, lighting, timing, and mood in a shot-by-shot prompt instead of working from a vague visual idea.
Use a first frame to preserve the subject, composition, brand treatment, or product appearance you have already signed off, and add a second image to guide the ending frame.
Generate compact 4–15 second clips for intros, product moments, explainers, social hooks, and short narrative beats.
Start at 768P for faster creative exploration or choose 2K when finer visual detail matters for a key asset.
See credit costs for MiniMax H3 and every other model on the pricing page. Credits never expire.
View all pricingMiniMax H3 is an AI video model that generates video from a text prompt, a single start image, or a pair of first and last frames. On ImagineGo, describe the scene and direction, optionally upload images, choose a duration of 4–15 seconds, and render at 768P or 2K.
Write the action as a sequence. Include the subject and setting first, then specify camera movement, lighting, timing, environmental motion, visual elements to preserve, and anything the model should avoid.
The first image anchors the opening subject, composition, and visual identity. A second image gives the generation a target end state, so your prompt can focus on the camera path and motion that connect the two frames.
Generate clips from 4 to 15 seconds at 768P or 2K. Supported aspect ratios include 21:9, 16:9, 4:3, 1:1, 3:4, and 9:16 for widescreen, square, portrait, and vertical placements.
The first five input images are included. The MiniMax H3 form accepts up to two frames, so a normal first-frame or first-and-last-frame generation does not add an image-input charge.
Generation is billed by output duration: 8 credits per second at 768P and 13 credits per second at 2K. For example, a 10-second 768P clip costs 80 credits.