fal H3 Max Turbo
#21 en Texte-vers-vidéoFal · max turbo · 2× · vu le 04 sept. 2026
36
Momentum
fal H3 Max Turbo is a text-to-video and image-to-video model released by fal (fal.ai), offered as a faster and more cost-effective variant of fal H3 Max. It is based on the Open Source MiniMax-H3 base model, Fine-Tuned by fal, and optimized for fal's proprietary Inference stack. According to fal, it runs at approximately double the speed and half the cost compared to H3 Max, while quality is expected to approach H3 Max closely (around the 97th percentile). The model is offered exclusively as a hosted API/Preview on fal.ai, not as a downloadable Open Source version.
Historique du momentum
08.06.06.09.
Fonctionnalités
| Fine-Tuning | No fine-tuning available; prompt-based control only (text-to-video, image-to-video with optional end frame), no reference-to-video support |
| Generation Time | Roughly 2x the speed of H3 Max (H3 Max: 5s clip at 768p in under 3s) – so well under 3s per 5s clip |
| License | No open-weight release; available only as a fal-hosted model ('everything but open sourcing the weights') |
| Max Resolution | 768p (480p also available) |
| Max Video Length | Up to 15 seconds per generation |
| Platform | Hosted API on fal.ai (text-to-video and image-to-video endpoints) |
| Price | Promo: $0.01/s at 768p (implies ~$0.02/s regular, half the price of H3 Max); promo runs for 2 weeks from announcement |
| Release Date | Preview announced around September 2, 2026 |