Home
Creators
Search Recent Random
Posts
Search DMs Popular Hash Lookup Tags Random
Importer
Import FAQ
Register Favorites Keys Review DMs Logout
Community
Telegram ThePornDude
Home Creators Posts Import Register Favorites Logout

Z-IMAGE TURBO NEW TEXT TO IMAGE AI KING WORKFLOW-V2! (Patreon)

Published:
2025-11-27 17:14:34
Edited:
2025-12-15 01:40:21
Imported:
2026-06
Tags:

Downloads

Content

EDIT 7 DECEMBER 2025: Hey everyone! I just released a new video update for the V2 Workflow that you can watch here: https://youtu.be/eyQkAW4wL1EIn this update I added a ControlNet workflow for better pose and composition control, a Detailer Daemon pass to crank the micro details even higher, AND THE AMAZING SeedEnhancer node that finally fixes on of the biggest issues with the model: the “different seed, same image” problem, a fantastic fix that allows you to generate much more different images with the same prompt! All tuned for low VRAM with Z-IMAGE TURBO.

Just use the Z-IMAGE-TURBO-MODELS-NODES_INSTALL-V2.bat installer from this post:https://www.patreon.com/posts/z-image-turbo-ai-144536371put it in your comfyui folder, run it to install the missing nodes and files, then update comfyui to the latest version with the update comfyui bat file inside the update folder then download the the updated workflow from this post drop it into your ComfyUI folder and you’re ready to go 😉Hey everyone! I’ve just released a new video on THE NEW TEXT TO IMAGE AI KING, called “Z-IMAGE TURBO”. It’s a completely insane 6B open source model that generates state of the art images, crazy prompt following and full unrestricted freedom and runs with LESS THAN 8GB OF VRAM! JUST INSANE! This is one of the biggest model announcement we've had in a long time!

You can check out the video right here: https://youtu.be/OtPd4ObjiDY

I also created a 1-click installer for the Webui for my Patreon supporters for Local AND Runpod, that you can find right here: https://www.patreon.com/posts/144536371

Here the link to the manual install guide: https://pastebin.com/RGSfk7j6

Down below you will also find the Z-IMAGE_TURBO_ULTRA_WORKFLOW.json for free for you to use!

As always, supporting me on Patreon allows me to keep creating helpful resources like this for the community. Thank you for your support - now go have some fun😉!

Files

Comments

Vyacheslav

Hi! Thanks for being so quick with this model. But I can't use it, because of getting this error: ClipLoaderGGUF unexpected text model architecture type in GGUF file: "Qwen3"

Ron Paul
>>194961982

I had the same error. Manager -> Update: ComfyUI stable version -> Update all

Sean Wilton

Hey I don't see the json file attached to this thread, Am I just not seeing it? lol

Hash

i see

Jaina Shop

same I don' t see it :(

Jim Sink

Anyone else getting really ugly skin textures when using inpainting?

Kurt C

Frick ya this aint easy. Wasted an hour trying to set it up on runpod. Was always missing something. Or was loading an old cache?? Will try again tonight.

Forest

How do I add checkpoints? I see the Lora are getting released already and one checkpoint.

Aitrepreneur

what do you mean checkpoints? There is only 1 Z-image turbo model right now, for other models you need different workflow

Forest

Thank you for your reply. It was my mistake, its the samething we have. Someone uploaded it on CivitAI and called it a checkpoint.

Stanley Mondong

hey i just tried to install 1 click etc. when i open comfy ui and drop .json files, it said "workflow has missing nodes" please help

Jason

Getting the missing nodes error now... was working fine before today

LW

If someone has released a modified version of Z-image Turbo at CivitAI, for example, and it is in GGUF format, then you can load it in to the same folder as the original model and then when you start ComfyUI it should appear in the pulldown menu for the load-model node.

Aitrepreneur

update comfyui using the bat file, and update the nodes from the manager.

Aitrepreneur

update comfyui using the bat file, and update the nodes from the manager.

Tom Baron

I am getting this error message when trying to run. CLIPLoaderGGUF Unexpected text model architecture type in GGUF file:'quen3'

Aitrepreneur

update comfyui with the bat file and update the nodes via the manager

Stanley Mondong

yes it's done thank you, but it only work for my rtx4070. does it not support gtx 1080?

Tom Baron

thank you, that solved it.

grstar85

The manual guide link does not work (nor the one on youtube)

Charles Leonhart

Is controlnet available for this model and can be added to the workflow?

Jason

Yeah already did that... no difference. Even did a fresh install with it all and nothing.

Aitrepreneur

yes a pseudo controlnet, it's not great but I can add it

grstar85

I mean there is no guide when I click the links. This jus fyi, I was able to install without the guide, but finding the models without a link was a pain.

Airsoft Fow

i get this error when i try to execute the workflow: Unexpected text Model architecture type in the GUFF file: 'qwen3'

Dannaman

Hi. When you say "Just use the Z-IMAGE-TURBO-MODELS-NODES_INSTALL-V2.bat installer from this post: *put it in your comfyui folder,* is that the "ComfyUI_windows_portable" or the "ComfyUI_windows_portable/ComfyUI" folder? There's a bit of ambiguity. Thanks.

Arthur Pewtie III

Any idea why when I try to run the control net option I get this error? ModelPatchLoader cannot access local variable 'model' where it is not associated with a value It seems associated with the model patch loader. The line around the model in the UI turns magenta, I have the model saved in the model_patches folder so that should be ok.

Aitrepreneur

I've seen that for some people you might have to update it to the comfyui nightly version for it, go inside your manager and choose the nightly version

Aitrepreneur

You also need to update it via the update comfyui bat file from the update folder, it's really just an update issue nothing else

Mojito69

Hi, just signed up as a member. This is the first of your workflow (WF) im trying out. Got a few questions: Your WF is using GGUF stuff... can I use non GGUF as well? For example, I already have Z-image-Turbo BF16 as well as qwen 3.4 clip loader.. If ok, how do I use it in your WF? the UNET LOADER is not listing my existing ZIT. The reason is, I am on 5070ti 16GB VRAM 64GB RAM btw, if that makes any difference

David Jeppesen

Working with this on Runpod. The workflow loaded just fine (thank you for that), but all of the Fast Groups Bypasser nodes are stuck with OPT_CONNECTION off and I cannot toggle any of them to on. Any ideas? Action taken: - Restarted ComfyUI - Opened in different browsers - Attempted to refresh/load missing custom nodes from the Manager but this always timed out without loading anything. Many thanks for any guidance you can provide!

llprimarinall

I just start using comfyui and when i use the workflow i receive alot of error such as, Failed to find the following ComfyRegistry list. The cache may be outdated, or the nodes may have been removed from ComfyRegistry. comfyui-aspect-ratio-crop-node comfyui-gguf also i can not find the the nodes that you used in this workflow in the missing nodes ??

llprimarinall

@aitrepreneur, it's work now with me but i received new error how could i solve this issue??? Prompt outputs failed validation: VAELoader: - Value not in list: vae_name: 'ae.safetensors' not in ['SDXL\\sdxl_vae.safetensors', 'taesd', 'taesdxl', 'taesd3', 'taef1', 'pixel_space'] CLIPLoaderGGUF: - Value not in list: clip_name: 'Qwen3-4B-UD-Q6_K_XL.gguf' not in [] UnetLoaderGGUF: - Value not in list: unet_name: 'z_image_turbo-Q8_0.gguf' not in []

Roni Bliss

Incredible! It's 2025 and there are still so many DA's who don't know how to solve ANY error within 10 seconds, instead flood every comments section wit "I have an error" never a thanks for the post! So for those who really should not be using anything close ro afvanced like Comfyui, and should go back to using MS Paint on Windows 95, here is a way to get an easy follow step by step guide on solving 99% of any errors, ComfUI or ANY other app in 2025. simply take those error logs you flood these comments sections with go to Google and instead of searching click "Ai mode" >> Paste the error log >> Click enter >> It will give you a step by step guide to fixing pretty much anything. To AIE thanks again for your excellent work! Best Z-image workflow out there. Depth Anythng V3 is an absolute game changer in Controlnet!

Simple Meanings

question why is the image generation so slow, running 16gb Vram and picked the 12-16gb version in the auto install

Maximilian Bauerle

I think he told in the video to take the small one. But rewatch it maybe. I have 10gb and it is fast

Patrick Kaiser

Hi, no matter what I try the Image to Image workflow doesn't follow any instructions. I uploaded my AI testimage of a woman, it could not place her in another setting, could not change/remove clothes. Just changed the face even if I explicitly wrote that the face and body shape should be exactly the same. Any ideas?

Aitrepreneur

it's not an edit model, image to image changes the whole image into another one depending on the denoise value

Patrick Kaiser

Thanks for your quick reply! I tested different denoise values, seeds and prompts but the output stays the same. It always just changes the body shape and the face while keeping the pose and the background the same for a prompt like this "Place this woman in a dimly lit bedroom, laying on the bed, looking at the camera. Keep the face and body shape exactly like in the input." and even the prompt "Style this image like a medival painting" has exactly the same effect. It seems like the prompt has no effect at all. Edit: The same problem occurs with inpainting. Here the picture just gets outputted without changes even for the example prompt "futuristic sunglasses"

Aitrepreneur

Man I just told you you are looking for an edit model to do something like that, wait for the z image edit model or watch my qwen image edit plus video

Kenneth Rosenstroem

Looking for the original first workflow, where do I find that?

Aitrepreneur

Why? The v2 is basically the same workflow as v1 but with added controlnet, seedvariance node and deamon detailer

Kenneth Rosenstroem

If that is the case, then yes, I don't need it. Thanks for your time and reply.

Avishek

I get a model mismatch error using the same models form sloth and jayn am i doing something wrong?