Video · Image → video
Hailuo 2.3
Turns an image into a 6- or 10-second clip — realistic motion and lifelike facial expressions
About the model
Hailuo 2.3 is MiniMax’s upgraded video model. Its strength is realism: natural body movement, believable scene physics, and subtle facial acting that keeps characters from looking stiff.
The model understands detailed motion and camera directions and keeps a consistent style from frame to frame — even in anime, watercolor, or game-grade CGI looks.
In our service, Hailuo 2.3 works in image-to-video mode and comes in two variants:
- Hailuo 2.3 Pro — for final output.
- Hailuo 2.3 Standard — lower cost, handy for testing ideas and drafts.
If you need video from a text description without a source image, take a look at Kling 2.6 or Seedance 2.0.
Strengths
- Realistic body movement and scene physics
- Subtle facial acting — emotion shifts, eye focus, micro-expressions
- Stable lighting, shadows, and color through complex poses and camera moves
- A wide range of styles — anime, illustration, watercolor and ink-wash, game-grade CGI
- Follows motion and camera directions precisely — pans, zooms, turns
- 768P resolution, 6- or 10-second clips
Best for
- Animating product renders, concept art, and key visuals
- Short promo clips and landing-page hero videos
- Character scenes where emotion and movement matter
- Stylized clips in anime or illustration looks
Limitations
- Image-to-video only — there is no text-to-video mode
- Input image size limit is 10 MB
- Supported formats: JPEG, PNG, WEBP
Prompting tips
- Attach a clean image — it becomes the first frame of the clip
- Describe motion and camera concretely — "pan to the right", "slow push-in", "jump and turn around"
- For close-ups, describe the character's emotion and gaze
- Standard costs less and suits tests; Pro is for final output