MiniMax H3 is MiniMax's flagship video model: 2K output, clips up to 15 seconds, and multimodal reference support. Generate from text, from an image, or from reference images and videos — all three cost the same.
Create a video now →| Model | Quality | Length | Audio | Image support | Price |
|---|---|---|---|---|---|
| MiniMax H3 | 768p / 2K | 4-15s | — | 5 images | from 320 credits ($3.20) |
From 4 to 15 seconds, at 768p or 2K. Pricing is per second, so short clips never cost more than they should.
Hailuo 2.3 is image-to-video only and tops out at 1080p. H3 also does text-to-video, outputs 2K, runs up to 15 seconds, and accepts image, video and audio references at the same time.
Yes — up to 3 reference videos totalling 15 seconds, to carry over motion and pacing. Requests with a reference video are priced on the input plus output seconds combined.
Protonin hosts 10 model families: Kling AI, Google Veo 3.1, Runway Gen-4, Wan 2.7, Hailuo AI, Seedance 2.0, Grok Imagine, Gemini video, HappyHorse AI.