MiniMax Hailuo 2.3
What is MiniMax Hailuo 2.3?
MiniMax Hailuo 2.3 is the auto-routing model in the Hailuo family. Pass a prompt, a start image, or both, and it picks between text-to-video and image-to-video automatically, so one call site handles every input mix. Output is 768p (default) or 1080p in 6 or 10 second videos; 1080p locks to 6 seconds. Prompts run up to 2000 characters, with an optional prompt_optimization rewriting pass.
Best for
Mixed-input upstream forms
One endpoint handles text alone, image alone, or both — no branching call sites.
Text-first storytelling
The 2000-character prompt budget carries narrative beats and detailed scene direction.
Image-to-video without routing logic
Add start_image_url and the model switches mode by itself.
When to use Hailuo 2.3
Pick Hailuo 2.3 when the upstream form lets users mix prompts and images and you want one model that routes itself. When the workflow is image-led and latency matters more, use Hailuo 2.3 Fast. For 2K output with native audio and reference control, step up to Hailuo 3.0.
API examples
Call Hailuo 2.3 by POSTing to /v1/tasks. Full parameter docs live at docs.unifically.com/models/video/minimax/hailuo.
curl -X POST https://api.unifically.com/v1/tasks \
-H "Content-Type: application/json" \
-H "Authorization: Bearer YOUR_API_KEY" \
-d '{
"model": "hailuo/minimax-2.3",
"input": {
"prompt": "A paper boat drifts down a rain gutter, camera tracking alongside at water level",
"duration": 10,
"resolution": "768p"
}
}'
Successful submission returns a task_id. Poll GET /v1/tasks/<task_id> or set a callback_url on the request to receive the finished result.
FAQs
People also ask
By input. Pass a prompt alone and it runs text-to-video; include start_image_url and it runs image-to-video. No mode parameter needed.
768p or 1080p, at 6 or 10 seconds. 1080p only supports 6-second videos.
2.3 accepts text-only input and treats the start image as optional. 2.3 Fast requires a start image, treats the prompt as optional, and returns faster.
