ComfyUI 0.32 runs LTX-2.5 and adds Grok Imagine Image 2.0
Release notesComfyUI 0.32.0 release notes on GitHub

ComfyUI 0.32.0 came out late on 11 August, the day Lightricks released LTX-2.5 — so the new model ran in ComfyUI from its first day.
LTX-2.5
The release adds support for LTX-2.5 for local generation, and LTX 2.5 partner nodes for running it on Lightricks' hosted service instead.
Partner nodes: Grok Imagine and Qwen
Two hosted image models arrive as partner nodes: xAI's Grok Imagine Image 2.0 and Qwen-Image 3.0 Pro — both useful as the source of a first frame for image-to-video.
MiniMax H3 gets lighter
MiniMax's H3 video model, opened the week before, gets an optimised VAE, a fix for VAEDecodeTiled with nested tensors, and a fix for its peak memory use. The release also fixes tiled audio decoding.
Before you update
The minimum PyTorch version is now 2.7. An older environment needs upgrading first, or ComfyUI will not start. The release also brings comfy kitchen attention and templates 0.11.39.
ComfyUI 0.32 runs LTX-2.5 and adds Grok Imagine Image 2.0: questions
Does ComfyUI support LTX-2.5?
Yes, from ComfyUI 0.32.0, released on 11 August 2026 — the day LTX-2.5 came out. It also has partner nodes for Lightricks' hosted LTX 2.5.
What PyTorch version does ComfyUI 0.32 need?
PyTorch 2.7 or newer.
Source: ComfyUI on GitHub checked against the source
More AI video news

LTX-2.5 is out, with open weights
Lightricks released LTX-2.5 on 11 August: a 22-billion-parameter open-weights AI video model that generates sound with the picture, cuts between shots in one clip and upscales to 3840×2176. It ran in ComfyUI on day one.

LTX-2.5 gets fourteen LoRAs and HDR
Lightricks spent September turning LTX-2.5 into a post-production kit: fourteen LoRAs for jobs like mattes, deblurring and day-for-night, long AI video in overlapping windows, and an SDR-to-HDR pass that writes EXR.

ComfyUI 0.38: Seedance drafts, H3 ControlNet
ComfyUI 0.38 brings Fun-ControlNet-Union 2.0 to MiniMax H3, a Draft mode for ByteDance's Seedance 2.5, film colour spaces for grading and a lighter install — and drops its Sora nodes now that OpenAI's AI video API is gone.