By GraphicCardPrices. Requirements checked against the official HunyuanVideo 1.5 documentation on .
This guide is for creators who want to turn a text prompt or a still image into a short video clip on their own PC, without paying per clip, and who want to know if a 16GB or 24GB card is enough.
HunyuanVideo 1.5 is Tencent's open video model. It has 8.3 billion parameters and does both text-to-video and image-to-video. The README lists 480p and 720p models and a super-resolution network that upscales the output to 1080p. The weights and inference code were released in November 2025.
License note: the weights use the Tencent Hunyuan Community License. It does not apply in the European Union, the United Kingdom or South Korea, and a company whose products have more than 100 million monthly active users must request a license from Tencent. Read the license file before using clips commercially.
What GPU do you need?
The official README says: "Minimum GPU Memory: 14 GB (with model offloading enabled)" (source)
| Tier | VRAM | What it runs | Example cards |
|---|---|---|---|
| Minimum | 14 GB | The README minimum, measured with model offloading on. Offloading keeps part of the model in system memory, so generation takes longer than on a bigger card. source | RTX 5060 Ti (16GB version), RTX 4060 Ti (16GB version), RTX 5070 Ti |
| Comfortable | 24 GB | The size ComfyUI's tutorial targets: it describes running the model "on consumer GPUs (24GB VRAM)". source | RTX 3090, RTX 4090 |
| Best | 32 GB | The most memory on a consumer card. This is our top tier rather than an official figure: the README gives no number for running without offloading. source | RTX 5090 |
Lowest recorded prices by tier
The cheapest New and Used listings we have recorded for each tier, among NVIDIA GeForce RTX cards with at least that much VRAM. Used includes renewed and refurbished cards. Each is a recorded Amazon US price, not a live checkout price; the model page shows every listing and its condition.
| Tier | Lowest New | Lowest Used |
|---|---|---|
| Minimum (14GB+) | $788.99 RTX 5060 Ti 16GB Recorded 2026-10-16 21:18:45 UTC | $699.97 RTX 4060 Ti 16GB Recorded 2026-09-29 16:18:54 UTC |
| Comfortable (24GB+) | $4,490.00 RTX 4090 24GB Recorded 2026-10-27 18:04:39 UTC | $1,699.99 RTX 3090 24GB Recorded 2026-09-29 16:27:49 UTC |
| Best (32GB+) | $7,199.99 RTX 5090 32GB Recorded 2026-10-16 21:16:50 UTC | No Used listing recorded |
How to get started
- Easiest route: install ComfyUI and follow the official ComfyUI HunyuanVideo 1.5 tutorial, which lists the model files and the folders they go in.
- Command-line route: the README asks for an NVIDIA GPU with CUDA and Python 3.10 or higher. Clone the repository, run
pip install -r requirements.txt, download the weights from Hugging Face and rungenerate.pywith your prompt and resolution. - The README says model offloading is on by default and suggests turning it off only if your GPU has memory to spare. If you have more than 14GB and still run out of memory, it suggests an allocator setting listed in its Usage section.
- Python route: HunyuanVideo 1.5 is also in Hugging Face Diffusers. The Diffusers page shows
enable_model_cpu_offload()andvae.enable_tiling()as the memory-saving settings in its example.
Full instructions are in the official HunyuanVideo 1.5 repository.
Things to check before you buy
NVIDIA, AMD and Intel
The README lists "NVIDIA GPU with CUDA support" as the hardware requirement and does not mention AMD or Intel cards. The price table above covers NVIDIA GeForce RTX cards. The README also says its sparse attention option requires H-series GPUs, so skip that option on a consumer card.
System RAM
Offloading moves work to your system memory. The README warns that its overlapped group offloading option "significantly increases CPU memory usage" and gives a flag to switch it off, but it states no RAM figure.
Speed
For the 480p image-to-video step-distilled model, the README states a single RTX 4090 can generate a video within 75 seconds. That is Tencent's own figure for one model and one card, not our measurement. We have not benchmarked HunyuanVideo, and offloading on a smaller card will take longer.
Under 14 GB?
The README links a community project, Wan2GP, which it describes as a low-VRAM app "as low 6 GB of VRAM for Hunyuan Video 1.5". It is not maintained by Tencent and gives no details in the README, so treat it as a lead to test, not a guarantee. If you are choosing between video models, see our Wan 2.2 guide.
HunyuanVideo 1.5 questions
How much VRAM does HunyuanVideo 1.5 need?
The official README lists a minimum of 14 GB of GPU memory with model offloading enabled. ComfyUI's tutorial targets a 24GB consumer GPU. The README gives no figure for running without offloading.
Can I run HunyuanVideo 1.5 on a 12GB or 8GB card?
Not by the README's own minimum of 14 GB. It links a community app, Wan2GP, that it says reaches as low as 6 GB, but that is a third-party claim we have not tested.
Can I sell videos made with HunyuanVideo 1.5?
The Tencent Hunyuan Community License does not cover the European Union, the United Kingdom or South Korea, and products with more than 100 million monthly active users need a separate license from Tencent. Read the license file in the official repository first.
Compare cards
Estimate memory for language models with the LLM VRAM estimator, see typical prices per memory size in GPU prices by VRAM, or compare pairs: RTX 5060 Ti vs RTX 4060 Ti, RTX 3090 vs RTX 4090. More local AI GPU guides.
Sources
- HunyuanVideo-1.5 README (14 GB minimum, offloading, install), checked
- HunyuanVideo 1.5 license, checked
- ComfyUI HunyuanVideo 1.5 tutorial (24GB target), checked
- Diffusers HunyuanVideo-1.5 page (offloading and tiling), checked
We have not benchmarked these tools. Requirements change between releases, so check the repository before you buy.