Model reference · open weights
Pika-H3-Flex-QAD is an open-weight video model from FastVideo. Pika-H3-Flex-NVFP4-QAD (NVFP4) weighs 106 GB; the smallest configuration that runs it is L40S 48 GB.
For self-hosting, compare this build's requirements below with the GPU you plan to rent.
Summary of the FastVideo/Pika-H3-Flex-NVFP4-QAD model card, 2026-11-04
What it is
| Released by | FastVideo |
|---|---|
| Released | 2026-11-01 |
| Parameters | 13.9B |
| VRAM | 106 GB for the weights |
What it runs on
| Card | Weights | Memory |
|---|---|---|
| RTX 3060 12 GB … RTX 5090 32 GB | does not fit | |
| L40S 48 GB | tight (encoders offloaded) | 44.0 GB |
| A100 80 GB | fits (encoders offloaded) | 78.2 GB |
| H100 80 GB | fits (encoders offloaded) | 78.1 GB |
| RTX PRO 6000 Blackwell 96 GB | fits (encoders offloaded) | 93.8 GB |
| DGX Spark (GB10) 128 GB unified | tight | 107 GB |
| H200 141 GB | fits | 138 GB |
| B200 180 GB | fits | 176 GB |
From the model card
The NVFP4 version of Pika H3 Flex for NVIDIA Blackwell GPUs. It generates video with synchronized audio from a prompt, from first and/or last frames, or from a mix of image, video and audio references. Both Flex adapters are included, and quantization-aware distillation (QAD) against the BF16 Flex model optimizes them for NVFP4 execution.
Pika H3 Flex · FastVideo · Inference guide · Training recipe · FP4-attention variant
| Task | Inputs | Transformer |
|---|---|---|
| T2VA | Prompt | transformer/ |
| I2VA, L2VA, FL2VA | Prompt plus a first frame, a last frame, or both | transformer/ |
| Ref2VA | Prompt plus up to 9 images, 3 videos and 3 audio clips | transformer_ref/ |
Install FastVideo with FlashAttention-4, then run:
FASTVIDEO_ATTENTION_BACKEND=FLASH_ATTN FASTVIDEO_FA4=1 \
python examples/inference/basic/basic_pika_h3_flex.py \
--model-path FastVideo/Pika-H3-Flex-NVFP4-QAD --nvfp4-dit h3_dit \
--nvfp4-decoder /vae_decoder_nvfp4/decoder-nvfp4.safetensors \
--task ref2va --reference image:character.png --reference video:motion.mp4 \
--prompt "$(cat prompt.txt)" --resolution 768p --duration 5
For keyframe tasks, use --task t2va, i2va, l2va or fl2va, adding --image (first frame) and/or --last-image as the task needs. The inference guide covers every task, canvas sizes and durations, attention backends, and the NVFP4 options.
| Component | Purpose |
|---|---|
transformer/ | NVFP4 transformer for T2VA, I2VA, L2VA and FL2VA |
transformer_ref/ | NVFP4 reference-conditioned transformer for Ref2VA |
text_encoder/ | NVFP4 Qwen3-VL text and vision encoder |
vae/, audio_vae/ | Video and audio VAEs |
vae_decoder_nvfp4/decoder-nvfp4.safetensors | NVFP4 video decoder |
scheduler/, audio_scheduler/, processor/, tokenizer/ | Pika H3 Flex sampling and input processing |
pika_h3_flex_merged.json | Marks that the Pika adapters are already merged into the transformers |
The transformers are packed NVFP4 checkpoints and load directly in FastVideo, with no conversion step. Each one already contains its Pika adapter, so FastVideo does not apply it again.
Compared with post-training quantization alone, QAD brings the model 14-17% closer to the BF16 teacher on held-out inputs, across both adapters. The full recipe, including calibration, training and export, is open source: see the training recipe.
| Model | Attention | Choose it for |
|---|---|---|
| Pika-H3-Flex-NVFP4-QAD (this model) | BF16 | Output closest to BF16 Flex |
| Pika-H3-Flex-NVFP4-FP4Attn-QAD | NVFP4 Q/K | Faster long clips |
This model is a derivative of MiniMax H3 and Pika H3 Flex. It is distributed under the MiniMax H3 Community License Agreement; see NOTICE, THIRD_PARTY_NOTICES.md and licenses/. The Qwen3-VL text encoder is licensed under Apache 2.0.
We thank Pika for open-sourcing Pika H3 Flex, and MiniMax for releasing MiniMax H3.
Quoted from the model card on Hugging Face. The full card is behind the Hugging Face link above.