Skip to content
AIBRAINWORX / OpenSourceCoder.comLOCAL-FIRST. BUILT IN THE OPEN.
Creative & 3D Tools / Multimodal & Vision / Creative & Media / Models & Weights

ComfyUI 0.45 adds MiniMax-H3-image support

The release also updates workflows and includes partner nodes that may rely on external services.

Listen to this articleBrowser Voice
Ready
ComfyUI GitHub repository artwork
Official publisher or project artwork; editorial coverage.
Event / context: ComfyUI v0.45.0 was released November 10, 2026.. Source check: Nov 12, 2026.

ComfyUI 0.45.0 adds MiniMax-H3-image support and updates parts of its visual generation workflow, according to the project's release notes. The release updates the frontend and templates and improves an error message for an incorrectly selected pixel-space VAE.

Check where each node runs

Changes include adaptive SeedVR2 VAE slice length based on available memory and integrations with outside partners. A partner node is not necessarily a local model: check whether it sends data to a provider, requires credentials or has separate costs.

The MiniMax-H3-image change adds application support; it does not mean weights are bundled or guarantee a workflow fits a 12 GB GPU. Check the model's original terms and hardware requirements.

Back up workflows and check custom-node compatibility before upgrading. The repository has installation details and source. We have not installed or benchmarked this release.

EXPLORE THE SOURCE

GitHub repositories

Original reporting based on the linked publisher sources; no independent product or benchmark test.

Follow your curiosity.

Pass the signal.