Skip to content
Agents tracked: 492 Releases (7d): 545 Agent status: 1 with issuesSince last limit reset: Claude 9d · Codex 7d For AI agents: Markdown Updated Nov 20, 2026

HunyuanVideo 1.5 vs Wan 2.2

Both are open-weight video models you run yourself, with ComfyUI and Diffusers support. HunyuanVideo 1.5 is a single 8.3B model that runs from 14 GB of GPU memory with offloading, but its Tencent license does not apply in the EU, UK or South Korea; Wan 2.2 is Apache 2.0 with no territory limits, and its biggest models need 80 GB GPUs.

Content asset vs content asset · checked against the official pages on Oct 25, 2026

Side by side

HunyuanVideo 1.5Wan 2.2
What it is8.3B-parameter text-to-video and image-to-video model from Tencent that runs on consumer GPUs.Open MoE video model family for text-to-video, image-to-video, speech-to-video and character animation at 480p/720p.
PriceFreeFree
Free tierOpen weights, free to download; not licensed in the EU, UK or South Korea.Open weights, free to download; self-hosted (the 5B model runs on a single RTX 4090).
LicenseTencent Hunyuan Community LicenseApache-2.0
Commercial useWith conditionsAllowed
How an agent connectsDownload onlyDownload only
GitHub stars4,60918.8K
Gid Score2540

From AgentGid's catalogs: prices and licenses as read on the official pages, downloads and stars from npm, PyPI and GitHub, "used in agent projects" from our scan of public repositories set up for AI coding agents. Each item's page has its sources and dates.

Key differences

  • HunyuanVideo 1.5HunyuanVideo 1.5 is an 8.3B-parameter video model aimed at consumer GPUs. Source
  • HunyuanVideo 1.5Tencent lists a minimum of 14 GB of GPU memory when model offloading is turned on. Source
  • HunyuanVideo 1.5The Tencent Hunyuan Community License does not apply in the European Union, the United Kingdom or South Korea. Source
  • HunyuanVideo 1.5Services with more than 100 million monthly active users must ask Tencent for a separate license. Source
  • Wan 2.2Wan 2.2 is licensed under Apache 2.0, and its team claims no rights over generated videos. Source
  • Wan 2.2Wan's 14B text- and image-to-video models need a GPU with at least 80 GB of VRAM in the single-GPU setup. Source
  • Wan 2.2The Wan 2.2 family goes beyond text and image to video, with speech-to-video and character animation models. Source

Which one to pick

Pick HunyuanVideo 1.5 if…Choose HunyuanVideo 1.5 when you want one mid-sized video model that runs on a single consumer GPU with offloading, and you operate outside the EU, UK and South Korea. Review the Tencent license terms before using it in a product.
Pick Wan 2.2 if…Choose Wan 2.2 when you need a permissive Apache 2.0 license that works worldwide, or extra tasks such as speech-to-video and character animation. Use its 5B model on a 24 GB card, or plan for 80 GB GPUs for the 14B models.

Sources

Vendor figures are as reported on their own pages; AgentGid did not test these products. Not legal advice: read the license before you ship.

Compare with another one → · All comparisons of libraries, APIs and assets

More comparisons

Apify vs Firecrawl · Browserbase vs Hyperbrowser · Browserbase vs Steel · Cartesia vs ElevenLabs · Clerk vs Auth0 · Deepgram vs AssemblyAI · E2B vs Daytona · ElevenLabs vs Kokoro-82M