One multimodal model for image, video, audio, and action prediction. FLUX 3 Video is in early access.