Model reference · open weights
FastMetal-QAD is an open-weight video model from FastVideo. FastMetal-14B-QAD (BF16) weighs 27.1 GB; the smallest configuration that runs it is RTX 4090 24 GB.
Summary of the FastVideo/FastMetal-14B-QAD model card, 2026-10-01
What it is
| Released by | FastVideo |
|---|---|
| Released | 2026-08-10 |
| VRAM | 27.1 GB for the weights |
What it runs on
| Card | Weights | Memory |
|---|---|---|
| RTX 3060 12 GB … RTX 4060 Ti 16 GB | does not fit | |
| RTX 3090 24 GB | fits (encoders offloaded) | 23.4 GB |
| RTX 4090 24 GB | fits (encoders offloaded) | 23.4 GB |
| RTX 5090 32 GB | tight | 31.0 GB |
| L40S 48 GB | fits | 44.0 GB |
| A100 80 GB | fits | 78.2 GB |
| H100 80 GB | fits | 78.1 GB |
| RTX PRO 6000 Blackwell 96 GB | fits | 93.8 GB |
| DGX Spark (GB10) 128 GB unified | fits | 107 GB |
| H200 141 GB | fits | 138 GB |
| B200 180 GB | fits | 176 GB |
From the model card
3-step text-to-video, INT8 pre-quantized for Apple Silicon.
FastMetal is a family of DMD2-distilled Wan video models with quantization-aware-trained INT8 DiTs. We ship the DiT already quantized, so there is no startup quantization — download, load, generate.
This repo is fully self-contained for generation:
| Path | Contents |
|---|---|
mlx_dit.safetensors / mlx_dit.json | INT8 (affine, group-64) DiT — student checkpoint |
ema/ | EMA-smoothed variant of the same DiT |
text_encoder/, vae/, tokenizer/, scheduler/ | everything needed to run standalone (fp16 UMT5 text encoder) |
Requires macOS with Apple silicon (MPS) and Python 3.11+:
pip install torch transformers mlx safetensors av imageio imageio-ffmpeg
git clone https://github.com/FastVideo/FastVideo.git
cd FastVideo
# student checkpoint
python examples/inference/basic/mlx_wan_prompt_to_video.py \
--model-root ./FastMetal-14B-QAD \
--mlx-checkpoint ./FastMetal-14B-QAD \
--prompt "a cinematic slow pan over a mountain river at golden hour"
# EMA-smoothed variant
python examples/inference/basic/mlx_wan_prompt_to_video.py \
--model-root ./FastMetal-14B-QAD \
--mlx-checkpoint ./FastMetal-14B-QAD/ema \
--prompt "a cinematic slow pan over a mountain river at golden hour"
| Base model | FastWan 2.1 T2V 14B |
| Distillation | DMD2, 3 denoising steps |
| Quantization | affine INT8, group size 64, QAT-trained |
| Resolution | 448×832 (480p), 77 frames |
| Flow shift | 8.0 |
| DiT weights | ~15 GB (INT8) |
DMD2 distillation of the FastWan 2.1 T2V 14B teacher onto an INT8 student
on NVIDIA GB200 clusters, with quantization-aware training (affine INT8,
group 64) so the deployed model matches the training objective.
Training corpus: FastVideo/Wan-Syn_77x448x832_600k.
| Model | Tier |
|---|---|
| FastMetal-1.3B-QAD | Entry — 16 GB+ class Macs |
| FastMetal-5B-QAD | Mid — 720p |
| FastMetal-14B-QAD | Quality — 24 GB+/ Ideally 36 Macs |
Quoted from the model card on Hugging Face. The full card is behind the Hugging Face link above.
How it works
Running it yourself
Rent a machine by the hour. ComfyUI is installed on it. Open ComfyUI through the tunnel and load the workflow from the model's card on Hugging Face; choose this model's file in its loader.
# on your rented machine (the ssh line is on its page in the console)
# get REPO FILE FOLDER: one file into /workspace/models/FOLDER, where ComfyUI loads it from
get() { hf download "$1" "$2" --local-dir /workspace/hf-files && mkdir -p "/workspace/models/$3" && mv "/workspace/hf-files/$2" "/workspace/models/$3/$4"; }
# the model (14.1 GB)
get FastVideo/FastMetal-14B-QAD mlx_dit.safetensors diffusion_models
start-comfyui# on your computer, in a second terminal: ComfyUI in your browser at http://localhost:8188
# HOST and PORT are your machine's, from its page in the console
ssh -L 8188:localhost:8188 dev@HOST -p PORT