Creative AI

ComfyUI v0.39.0 Release Updates

The latest release of ComfyUI, a GitHub project, brings several updates and improvements, including support for LynnReal light Minimax-H3 vae and updated embedded docs.

2 min read375 words
AI-generated editorial illustration for: ComfyUI v0.39.0 Release Updates
Illustration: AI-generated for this story.

The latest release of ComfyUI, a GitHub project, brings several updates and improvements to the table. At the core of these updates is support for LynnReal light Minimax-H3 vae, a feature added thanks to the efforts of @kijai in issue #16657. This support enables users to leverage the power of Minimax-H3 vae in their applications, making it a valuable addition to the ComfyUI ecosystem. In addition to this major update, ComfyUI's embedded docs have been updated to version 0.5.13 by @comfyui-wiki in issue #16618. This update ensures that users have access to the latest documentation and information, making it easier to get started with ComfyUI. The release also includes several quality-of-life improvements, such as improved Save Video encoding defaults, a result of the efforts of @comfyanonymous. This update ensures that users can work with their media files with greater ease and confidence. Another key addition is the DynamicGroup widget input, added by @jaeone94 in issue #16260. This feature enables users to work with dynamic groups of assets, making it easier to manage and manipulate their media. The release also includes several bug fixes and performance improvements. One notable fix is the resolution of an issue that prevented model categories from aborting the scan when assets are being processed. This fix, implemented by an unknown contributor in an unspecified issue, ensures that users can work with their models without encountering errors. Furthermore, the release includes several updates related to partner nodes. The Sonnet 5.5 model, added by @bigcat88 in issue #16647, is now available for use, as is the Ideogram 4.5 model, also added by @bigcat88 in issue #16689. These models enable users to work with a wider range of assets and applications. Finally, the release includes a fix for a minimax vae offload issue, resolved by @comfyanonymous in issue #16698. This fix ensures that users can work with their minimax vae models without encountering errors. In conclusion, the latest release of ComfyUI brings several valuable updates and improvements to the table. From support for LynnReal light Minimax-H3 vae to updated embedded docs, this release has something for everyone. Tags: language models, open weights, regulation ImagePrompt: An illustration of a person working with multiple media files on a computer, with the ComfyUI logo in the corner.

Sources

Summarized from the linked originals. Cover illustration: AI-generated.

Related stories