HunyuanVideo-1.5
Tencent-Hunyuan/HunyuanVideo-1.5Reviewed by AI
Checked by RepoPicks Radar on
What is HunyuanVideo-1.5?
HunyuanVideo-1.5 is the lighter version of Tencent's video model: 8.3 billion parameters instead of 13. You give it a text description, or a photo plus a description, and it writes out a video at 480p or 720p, with an optional step that upscales to 1080p. It runs from the command line on Linux with an NVIDIA card; the authors state a minimum of 14 GB of video memory with offloading on. Model files are downloaded separately, and the project also documents use through ComfyUI and Diffusers.
HunyuanVideo-1.5 is free and can replace Runway, Kling, and Pika.
How easy is it to install HunyuanVideo-1.5?
D · terminal, several tools or a specific graphics card · Linux
Linux is the only system listed: you clone the repository, pip install the requirements, optionally build extra attention libraries from source, download the chosen checkpoints from Hugging Face, and run generate.py with a long list of arguments. An NVIDIA card with at least 14 GB of video memory is required.
Coming soon: install in 1 click with the RepoPicks app.
HunyuanVideo-1.5 on GitHub
- Stars
- 4,598on Nov 10, 2026
- Downloads
- —all releases
- Growth
- +7stars in 6 days
Stars over time
Facts
- Runs on
- Linux
- You need to install
- Python, NVIDIA graphics card, FFmpeg
- To build from source
- Python, Graphics card, FFmpeg
- Disk space
- 950 MB–4.3 GB · estimate
- Accounts or keys
- Prompt-rewriting model (Gemini or a self-hosted vLLM server)
- Cost
- Free
- License
- NOASSERTION (Other)
- Can replace
- Runway, Kling, Pika, LTX Studio, Veo 3
- Last change on GitHub
- Apr 10, 2026
HunyuanVideo-1.5 risk signals (5)
- AttentionDependencies with known vulnerabilities
Pinned dependencies with known vulnerabilities (OSV): pillow: GHSA-22hg-r758-rrw9, GHSA-24rc-8hj2-v98c, GHSA-254f-3936-288h; diffusers: GHSA-7wx4-6vff-v64p, GHSA-98h9-4798-4q5v, PYSEC-2026-2446; transformers: GHSA-29pf-2h5f-8g72, GHSA-69w3-r845-3855, GHSA-fgcw-684q-jj6r
automatic check
- Attentionlicense excludes eu uk south korea
LICENSE, first line: "THIS LICENSE AGREEMENT DOES NOT APPLY IN THE EUROPEAN UNION, UNITED KINGDOM AND SOUTH KOREA AND IS EXPRESSLY LIMITED TO THE TERRITORY"; the grant in section 2 is "for the Territory only", which excludes those regions.
AI review
- Attentioninstall step builds code from a personal fork
The SageAttention step in the README installs from an individual's fork rather than the original project: "git clone https://github.com/cooper1637/SageAttention.git ... python3 setup.py install".
AI review
- Infoinstall uses tencent package mirror
Step 2 of the install runs "pip install -i https://mirrors.tencent.com/pypi/simple/ --upgrade tencentcloud-sdk-python", pulling a Tencent Cloud SDK from a package index other than PyPI, with no explanation of what it is used for.
AI review
- Infolicense extra commercial conditions
The Tencent Hunyuan Community License is not a standard open-source licence: section 4 requires a separate licence from Tencent above 100 million monthly active users, and section 3(e) requires anyone offering a service built on it to state prominently that Tencent does not endorse it.
AI review
Evidence is shown in its original language.
What Radar could not check about HunyuanVideo-1.5
- How many gigabytes the checkpoints take, since you pick among a dozen model variants listed in the README
- Whether the stated 14 GB minimum really holds in practice, and how long one video takes on a single consumer card
- Whether the install works on Windows or macOS, which the authors do not list
- How much quality is lost without the prompt-rewriting step, which is on by default and needs a separate large model