FastH3 Omni — BF16 Release

This is the final BF16 release checkpoint of FastH3 Omni. It generates video with synchronized audio from a text prompt and ordered image, video and audio references, using eight transformer evaluations.

All release variants · UniServe · FastVideo

Choose a runtime

Hardware	Runtime and setup
NVIDIA GB200 / B200	UniServe deployment guide
NVIDIA RTX PRO 6000 Blackwell	UniServe deployment guide
NVIDIA Hopper, including H100 / H200	UniServe requirements and setup
NVIDIA DGX Spark (GB10) / RTX 5090	FastVideo local FastH3 Omni guide with the NVFP4 QAD variant

For server deployment, use this BF16 checkpoint with UniServe's reference-conditioned (ref2va) configurations. The guide covers installation, model downloads, GPU placement, request capacities and the video/audio API. For the BF16 deployment, choose UniServe's quality precision preset. NVFP4 presets require Blackwell hardware.

Local AI on DGX Spark and RTX 5090

Use FastVideo with the NVFP4 QAD variant for local inference. QAD means quantization-aware distillation; these weights are trained for NVFP4 execution on Blackwell. Start with the FastVideo cookbook and FastH3 Omni recipe builder.

Device	Setup and inference recipe
DGX Spark (GB10, 128 GB unified memory)	Install FastVideo · `spark` recipe
RTX 5090 (32 GB)	Install FastVideo · `rtx-5090` recipe

Both profiles default to 480p. Spark keeps the model components resident; the RTX 5090 offloads the text encoder and requires a precomputed AdaLN table. Follow the local inference guide to compose the model directory, prepare the table and choose a decoder. Keep the device recipe's attention and precision settings.

For FP8, see the FastH3 Omni FP8 weights and FP8 device setup, including the RTX 4090 recipe for hardware without FP4 support. Its bundled NVFP4 text encoder requires Blackwell; non-FP4 hardware needs compatible BF16 encoder components. FP8 and NVFP4 QAD outputs can differ from BF16.

The optional NVFP4 VAE decoders offer 36-layer quality and 8-layer fast tiers. The guide documents decoder selection with --decoder. These local variants have their own memory and precision settings.

For visual workflows, use the ComfyUI variants.

Inputs and outputs

Inputs: a text prompt plus up to nine images, three videos and three audio clips, with at most twelve references in total. At least one image or video is required.
Reference order: use <Picture 1>, <Video 1> and <Audio 1> tags to refer to the ordered inputs. Video references include their soundtracks.
Output: video with synchronized 32-kHz stereo audio at 24 FPS. Supported request sizes and lengths depend on the runtime and available memory.
Sampling: eight transformer evaluations, guidance 1.0. The supplied configuration sets the video/audio schedule and reference-conditioned attention.

Download

hf download FastVideo/FastVideo-FastH3-Omni-8-Step-V1 \
  --local-dir ./FastH3-Omni-BF16

The repository includes the BF16 reference transformer, schedulers and inference configuration. The runtime loads the shared text encoder, tokenizer, processor and video/audio VAEs from MiniMax-H3. Follow the runtime guide to prepare those components.

Acknowledgements

FastH3 Omni was built in close collaboration with Nuva Lab, which co-developed the model with FastVideo, defined and curated the omni reference dataset, and brought production-grounded creative video workflows to FastH3. For dedicated production deployment, fine-tuning on your own data and a creative agent on top, see Nuva Lab's FastH3-Omni page.

Thanks to Reactor for providing compute, and to the NVIDIA FastGen team.