Preview page for MiniMax H3. Currently executes through Grok Imagine Video 1.5 (default, no audio) in text-image-to-video mode. Generate a character video with one visible subject and action. Other selectable models may support audio.
Generate video · 6 credits total
Use this preview to generate a character video via Grok Imagine Video 1.5 (default, no audio). Write a prompt with one visible subject, one deliberate action, and a camera cue. Other selectable models may support synchronized audio.
Describe one character, one action, and one camera movement. Keep the scene simple to maintain consistency.
Watch the output for subject consistency, action clarity, and visual quality. Regenerate if needed.
Adjust the action, camera cue, or setting to refine the result. Each generation is independent.
A muscular blacksmith in a leather apron strikes a glowing sword on an anvil. Sparks fly with each hammer blow. Camera holds a medium shot. Ends on the glowing blade.
Confirm the blacksmith, anvil, and sparks are visible and consistent.
Describe 'sparks flying with each hammer blow' more vividly in your prompt.
Yes, write 'static camera, medium shot' for a fixed frame.
Specify 'ends on the glowing blade' as the final frame description.
