We've added MiniMax Hailuo 03 (H3) to video generation — MiniMax’s frontier video model with native stereo audio and 2K output:
- Hailuo 03 (H3): text-to-video and image-to-video (optional last frame)
- Hailuo 03 Reference: reference-to-video from images, videos, and audio
Announcement: MiniMax H3
Why This Matters
- Frontier quality at 2K: H3 renders at 2K with native stereo audio on every clip.
- Flexible durations: Generate 5–15 second videos (default 5s).
- Multimodal reference: Keep subjects consistent with up to 9 images, 3 videos, and 3 audio clips (≤12 files total).
- First/last frame control: Image-to-video accepts an optional end frame for guided transitions.
What's New
- Hailuo 03 (H3) — text-to-video and image-to-video, available in the video model selector under Minimax Hailuo
- Hailuo 03 Reference — reference-to-video from images, videos, and audio
- Existing Hailuo models (
hailuo-02,hailuo-2.3,hailuo-2.3-fast) remain available - Aspect ratios (text-to-video / reference-to-video):
16:9,21:9,4:3,1:1,3:4,9:16(reference-to-video also supports adaptive) - Resolution:
2Konly for now
Pricing
- Base (all modes): $0.338 / second @ 2K
- Reference extras: first 5 reference images free; each additional image $0.104; reference video billed at the same per-second rate; audio references free