Skip to content
InvokeAI Logo InvokeAI Documentation

Releases

InvokeAI 6.13.0 Latest

This release of InvokeAI offers support for a variety of new models, including those that are remotely hosted by external providers such as GPT Image. In addition, it provides a host of new and improved features as well as bug fixes.

Installation

See Installation for installation instructions and troubleshooting.

What’s New

New Locally-Hosted Models

We’ve added support for the popular Qwen Image models, both the txt2img and edit versions, as well as Anima.

Qwen Image 2.5

This is a high quality txt2img model that can be run either on your local hardware are on a remotely hosted service (see Hosted Models. We provide starters for the full model (~40 GB!) as well as multiple smaller quantized models. For best results, we recommend 30-40 steps and a CFG of 3-4. We also provide two “turbo” LoRA models, Qwen Image Lightning (4-steps) and Qwen_Image_Lightning (8-steps). These LoRAs provide fast rendering and high accuracy, but loss of diversity. Use a CFG of 1 and a Shift of 3 when applying either of these LoRAs.

Qwen Image Edit 2.5

This is a version of Qwen Image that is optimized for image editing. Add up to three reference images and tell the model on how to combine and alter them using declarative statements such as “Pose the duck on the table”. We provide the full model as a starter, along with multiple quantized versions and two Qwen Image Edit Lightning LoRAs for turbo mode.

See Qwen-Image-Edit: Image Editing with Higher Quality and Efficiency for tips on using Qwen Image Edit effectively.

Anima Preview 3

Anima is an anime-themed model created by a collaboration between CircleStone Labs and Comfy Org. It is able to produce a wide variety of illustration styles. There is also an Anima Turbo LoRA that can be used to speed up rendering using a CFG of 1 and 8-12 steps.

You’ll find Anima and its support models in the Starter Models tab.

New Externally-Hosted Models

This version of InvokeAI offers the ability to generate using externally-hosted models, such as those provided by OpenAI, Google and cloud-based AI generation hosting services. You “install” these models by providing an API key provided by a service that you subscribe to. Typically image generation services will provide a number of generations for free and then start charging on a pay-as-you-go basis. We support externally-hosted models from OpenAI, Gemini (Google), BytePlus, and Alibaba Cloud.

An externally-hosted model works much like a locally-hosted one but with restrictions. In general, you cannot adust steps, scheduler or CFG, and only a limited number of image dimensions are supported. Both txt2img and img2img modes are supported, and the models can be applied to the Canvas and Workflow editor. However inpainting, outpainting, controlnets, LoRAs, regional guidance, and the ability to apply different levels of denoising to img2img are not available. These are limitations imposed by the hosting service and not by InvokeAI.

To get started with externally-hosted models, go to the Models tab and select External Providers. Select the hosting service you wish to use and enter your API key. Click Save and the models supported by this provider will be added to your collection of models.

Models available from OpenAI

  • GPT Image 1
  • GPT Image 1.5
  • GPT Image 1 Mini
  • DALL-E 3

Models available from Gemini (Google)

  • Gemini 2.5 Flash Image
  • Gemini 3 Pro Image Preview
  • Gemini 3.1 Flash Image Preview

Models available from BytePlus

  • Seedream 3.0 T2I
  • Seedream 4.0
  • Seedream 4.5
  • Seedream 5.0 Lite

Models available from Alibaba Cloud

  • Qwen Image 2.0 Pro
  • Qwen Image 2.0
  • Qwen Image Max
  • Qwen Image Edit Max
  • Wan 2.6 Text-to-Image

Model Manager Improvements

This version improves the Model Manager in several ways. When installing a new model, the installation queue display and progress bars have been completely rewritten to be more functional and attractive. A new bulk reidentify action has been implemented, and it is now possible to sort the model listing in multiple ways.

Image Generation Improvements

Users now have access to LLM-powered prompt expansion and image-to-prompt features. You will find two new buttons in the prompt textfield:

  • Expand Prompt uses a local TextLLM model to expand brief prompts into detailed image generation prompts. This button will replace a simple prompt with a complex one that is well-suited for producing high quality imagery.
  • Image to Prompt Generates descriptive prompts from uploaded images. Use it as a supplement to reference images.

To use these features, you must first install a causal LM model, such as Qwen/Qwen2.5-1.5B-Instruct via the Model Manager (enter this repo name into the HuggingFace tab). The LLaVA vision model needed for image to prompt should already be installed.

In addition, this release corrects many small bugs and performance bottlenecks in image generation.

You can now automatically create “virtual” boards that group images according to their characteristics. The only grouping currently available is by date, but more options are on their way.

Canvas Improvements

There have been several new features added to the Canvas. One frequently-requested feature is the ability to hide and show the preview stage thumbnails that can obscure the bottom of the canvas. Another feature is the addition of a Lasso Tool with freehand and polygon modes for creating inpaint masks and selecting raster image regions.

Also among multiple new features is the ability to apply custom workflows directly to a Canvas image. First create a workflow with an image input field in the Form Builder and a Canvas Output node. Right click on the canvas and select Run Workflow. The workflow will run on the input image and create a new layer containing the output of the Canvas Output node.

Workflow Editor and Nodes

There is now a Custom Node Manager tab in the left-hand panel which allows you to install and manage related groups of nodes organized into node packs. Node packs can be installed from local files or URLs.

When adding a node to a workflow, you will find that the nodes are now organized by function with an easy to navigate node picker. Another convenient new feature is the workflow connector, a new type of edge that can be used to connect two or more input/output fields in a highly flexible fashion. To start using this feature, right click on the canvas and select Add Connector.

Internally the workflow execution engine has been overhauled and optimized to improve performance and stability. In addition, there are a number of new user-visible features. There are new generation nodes for running external models, new denoising nodes for the Qwen Image and Anima local models, and a node for decoding invisible watermarks in images. The Collect node now supports chaining to enable deterministic iteration over the items in a collection. A new If mode allows for conditional branching of workflows.

Multiuser Features

When running in Multiuser Mode each logged-in user can create private (only visible to the owner), shared (writable by the owner, readable by everyone else) and public (read/write for everyone) image boards. Similarly, users can now create private and shared workflows.

Documentation

The Documentation Site has been completely redesigned and overhauled.

What’s Changed in Detail

Below is a log of all changes from the previous version. Many thanks to all the authors who contributed to this release.

Models and Model Manager

Image Generation

Image Gallery & Image Storage

Canvas

Workflow Editor and Nodes

Multiuser

Internationalizaton

Thanks to Dmitry Warkentin, Gohsuke Shimada, and Riccardo Giovanetti (@Harvester62) for contributing Russian, Japanese, and Italian translations, respectively.

Documentation

Miscellaneous

New Contributors

Full Changelog: https://github.com/invoke-ai/InvokeAI/compare/v6.8.0…v6.13.0

v6.12.0

InvokeAI v6.12.0

This is a feature release of InvokeAI which provides support for multiple accounts on the same InvokeAI backend, enhanced support for the Z-Image and FLUX.2 models, multiple user interface enhancements, and new utilities for managing models.

[Jump to Installing and Updating]

Multi-User Mode (Experimental)

Have you ever wished you could share your InvokeAI instance with your friends, family or coworkers, but didn’t want to share your galleries or give everyone the ability to add and delete models? Now you can. InvokeAI 6.12 introduces an experimental multi-user mode that allows you to create separate user accounts with login names and passwords. Each account’s image boards, images, canvas state and UI preferences are separate from the others. Users with administrative privileges are allowed to perform system-wide tasks such as adding and configuring models and managing the session queue, while ordinary users are prevented from making this type of change.

InvokeAI Log-In Screen

See the Multi-User Mode User’s Guide for information on setting up and using this mode.

Multi-User mode was contributed by @lstein .

Enhanced Support for Z-Image and FLUX.2 Models

Z-Image Base — This version of InvokeAI adds support for the Z-Image Base model family. This is an undistilled version of Z-Image suitable for fine-tuning and LoRA training. It also provides a high level of image diversity while preserving excellent image quality.

FLUX.2 LoRAs — InvokeAI now supports a variety of FLUX.2 Klein LoRA formats.

Thanks to @Pfannkuchensack for his work on these enhancements.

Paged Gallery Browsing — Paged gallery browsing is back. Go to image board settings and select “Use Paged Gallery View” to replace infinite gallery scrolling with page-by-page navigation.

image

Arrow Key Navigation — The arrow keys now work correctly when browsing a gallery. When the Viewer is in focus, the right and left arrow keys will navigate through the currently selected gallery. When the gallery thumbnails are in focus, the right/left/up/down arrows navigate among them.

@DustyShoe contributed these enhancements.

New Canvas Features

The Canvas now features several new features added by @DustyShoe

Text Tool — The Canvas now features a Text tool that allows you to insert text in a variety of fonts, sizes and styles, move it around the canvas, and commit it to the raster layer.

Linear and radial gradient tools — These new tools add radial and linear gradients to the Canvas. The gradients use color transparency and the foreground/background colors to draw gradients in the direction of the mouse movement.

image

Invert Button for Regional Guidance Layers — You can now select any Regional Guidance region and select the “invert” button to exchange painted regions with unpainted ones and vice versa. As an added bonus, the invert button also works with Inpaint Masks.

Layer Controls Moved The controls for creating, duplicating and deleting canvas layers have been moved from the top of the layers list to the bottom, which is more consistent with how other graphics packages position their layer controls and, we think, more intuitive. Long-term Canvas users may need to adjust to the new positioning.

A few improvements contributed by @lstein aim to make it easier to maintain the model and image databases.

Remove Orphaned Models — Over time InvokeAI may accumulate unused “orphan” models in its models directory that take up space but have no entries in the models database for one reason or another. This means they take up disk space without being usable. A new “Sync Models” button in the Model Manager detects and offers to delete such orphaned models. Developers and other users who have access to the source code repository will also find a script, located in scripts/remove_orphaned_models.py , that will do the same thing from the command line.

Remove Dangling Models — The converse problem occurs when a model directory, or one of its files, was removed or renamed externally, causing it to be referenced in the models database but not be usable. There is now a “Missing Files” filter option in the Model Manager that will identify models that are damaged or deleted. You can then select the models you wish to delete and remove them from the database. In addition, the model selection menus will no longer display models that are missing or broken.

Gallery Maintenance Script — For users with access to the source code repository, the scripts/gallery_maintenance.py python script will clean up dangling and orphaned gallery images. Dangling images are those that appear in the Invoke gallery database but whose files have been deleted from disk. Orphaned images are those that have files on disk but are missing from the database. A related database maintenance tool with more bells and whistles can also be found in @Pfannkuchensack ‘s GitHub at https://github.com/Pfannkuchensack/sqlite_invokeai_db_tool.

Workflow Iterator Improvements

@JPPhoto fixed the way that workflow collections work. Previously when you created a Collection and passed it to an iterator, the items in the collection would be passed to downstream nodes in an unpredictable order. Now, the order of items in the collection is preserved, making complex workflows more predictable and reproducible.

Remote Controlling Invoke’s Generation Parameters

It is now possible to programmatically set Invoke’s generation parameters using a new REST endpoint. This allows a script or other external program to select the model, image size, seed, steps, LoRAs, reference images, and all the other parameters that go into a generation. For documentation of the feature see:

@lstein added this feature.

Translations

Thanks to @Harvester62 for providing the Italian translations for this release.

Installing and Updating

The Invoke Launcher is the recommended way to install, update and run Invoke. It takes care of a lot of details for you - like installing the right version of python - and runs Invoke as a desktop application.

Note: With recent updates to torch, users on older GPUs (20xx and 10xx series) will likely run into issues with installing/updating. We are still evaluating how we can support older GPUs, but in the meantime users have found success manually downgrading torch. Head over to discord if you need help.

Follow the Quick Start guide to get started with the launcher.

If you don’t want to use the launcher, or need a headless install, you can follow the manual install guide.

Behind-the-Scenes Improvements

This release are contains a number of bug fixes and performance enhancements.

  • Optimize cache locking in Klein text encoder — (@girlyoulookthebest) This addresses a race condition in the Model Cache which prematurely removed the FLUX.2 Klein encoder from memory.
  • Run Text Encoder on CPU — (@lstein) This is an option available in the details panel of the Model Manager that allows you to force large text encoder models to run on CPU rather than GPU. This preserves VRAM for use by the denoiser steps and in some cases improves performance. Thanks to @girlyoulookthebest who found and fixed a bug in this feature.
  • Fix IP Adapters losing their model path — (@Pfannkuchensack) Fixes the Model Manager’s “reidentify” function when run on IP Adapter models.
  • Kill the server with a single ^C — (@lstein) When previous version of Invoke were launched from a command-line terminal, it used to require two key board interrupts (control-C) to completely shut it down. This is now fixed.
  • Persist the selected board and image across browser sessions — (@lstein) The last image board selected is now restored when you edit a browser session and restart it.

Detailed Change Log

What’s Changed

New Contributors

Full Changelog: https://github.com/invoke-ai/InvokeAI/compare/v6.11.0…v6.12.0

v6.11.1

InvokeAI 6.11.1

This is a bugfix release that corrects several image generation and user interface glitches:

  • Fix FLUX.2 Klein image generation quality (@Pfannkuchensack)
    • At higher step values and larger images, the FLUX.2 Klein models were generating image artifacts characterized by diagonals, cross-hatching and dust. This bug is now corrected.
  • Restore denoising strength for outpaint mode (@Pfannkuchensack)
    • Previously, when outpainting, the denoising strength was pinned at 1.0 rather than observing the value set by the user.
  • Only show FLUX.1 VAEs when a FLUX.1 main model is selected (@Pfannkuchensack)
    • This fix prevents the user from inadvertently selecting a FLUX.2 VAE when generating with FLUX.1.
  • Reset ZiT seed variance toggle when recalling images without that metadata (@Pfannkuchensack)
    • When remixing an image generated by Z-Image Turbo, the setting of the seed variance toggle (which increases image diversity) is now correctly restored.
  • Improve DyPE area calculation (@JPPhoto)
    • DyPE increases the quality of FLUX.1 models at higher resolutions.. This fix improves how the algorithm’s parameters are automatically adjusted for image size.
  • Remove duplicate DyPE preset dropdown in generation settings (@Pfannkuchensack
    • The DyPE dropdown in generation settings is no longer duplicated in the generation UI.

In addition to these bug fixes, new Russian translations were added by (@DustyShoe).

Checkout the roadmap

To see what the development team has planned for forthcoming releases, check out the InvokeAI roadmap. Feature releases will be issued roughly monthly.

Take the user survey

And don’t forget to tell us who you are, what features you use, and what features you most want to see included in future releases. Take the InvokeAI 2026 User Engagement Survey and share your thoughts!

Credits

In addition to the authors of these bug fixes, many thanks to @blessedcoolant, @skunkworxdark, and @mickr777 for their time and patience testing and reviewing the code.

Full Changelog: https://github.com/invoke-ai/InvokeAI/compare/v6.11.0…v6.11.1

v6.11.0

InvokeAI v6.11.0

This is a feature release of InvokeAI which provides support for the new FLUX.2 Klein image generation and edit models as well as a few small improvements and bug fixes. Before we get to the details, consider taking our 2026 User Engagement Survey. We want to know who you are, how you use InvokeAI, and what new features we can add to make the software even better.

Support for FLUX.2 Klein models

The FLUX2 Klein family of models (F2K) are fast, high quality image generation and editing models. Invoke provides support for multiple versions, including both the fast-but-less-precise 4 billion (4B) and the slower-but-more-accurate 9 billion (9B) models, as well as quantized versions of these models suited for systems with limited VRAM. These models are small and fast; the fastest can render images in seconds with just four steps.

In addition to the usual features (txt2img, img2img, inpainting, outpainting) F2K offers a unique image editing feature which allows you to make targeted modifications to an image or set of images using prompts like “Change the goblet in the king’s right hand from silver to gold,” or “Transfer the style from image 1 to image 2”.

Suggested hardware requirements are:

FLUX.2 Klein 4B - 1024×1024

  • GPU: Nvidia 30xx series or later, 12GB+ VRAM (e.g. RTX 3090, RTX 4070). FP8 version works with 8GB+ VRAM.*
  • Memory: At least 16GB RAM.
  • Disk: 10GB for base installation plus 20GB for models (Diffusers format with encoder).

FLUX.2 Klein 9B - 1024×1024

  • GPU: Nvidia 40xx series, 24GB+ VRAM (e.g. RTX 4090). FP8 version works with 12GB+ VRAM.
  • Memory: At least 32GB RAM.
  • Disk: 10GB for base installation plus 40GB for models (Diffusers format with encoder).

Getting Started with F2K

After updating InvokeAI, you will find a new FLUX.2 Klein starter pack for in the Starter Models section of the Model Manager. This will download three files: the Q4 quantized version of F2K 4B, which is suitable to run on low-end hardware, and two supporting files: the FLUX.2 VAE, and a quantized version of the FLUX.2 Qwen3 text encoder.

After installing the bundle, select the “FLUX.2 Klein 4B (GGUF Q4)” model in theGeneration section of Invoke’s left panel. Also go to the Advanced section at the bottom of the panel and select the F2K VAE and text encoder models that were installed with the starter bundle. (If you don’t select these, you will get an warning message on the first generation that tells you to do this.) Recommended generation settings are:

  • Steps: 4-6
  • CFG: 1-2

Modestly increasing the number of steps may increase accuracy somewhat. If you work with the Base versions of F2K (available from HuggingFace), increase the steps to >20 and the CFG to 3.5-5.0.

Text2img, img2img, inpainting and outpainting will all work as usual. InvokeAI does not currently support F2K LoRAs or ControlNets (there have not been many published so far). In addition, only the Euler sampler is currently available. Support for LoRAs and additional schedulers will be added in a future release.

Prompting with FLUX.2

Like ZiT, F2K’s text encoder works best when you provide it with long prose prompts that follow the framework Subject + Setting + Details + Lighting + Atmosphere. For example: “An elderly king is standing on a low dais in front of a crowded and chaotic banquet hall bursting with courtiers and noblemen. He is shown in profile, facing his noblemen, holding high a jeweled chalice of wine to toast the unification of his fifedoms. This is a cinematic shot set that conveys historical grandeur and a medieval vibe.”

F2K does not perform any form of prompt enhancement, so what you write is what the model sees. See FLUX.2 Prompting Guide for more guidance.

Image Editing

F2K provides an image editing mode that works like a souped-up version of Image Prompt (IP) Adapters. Drag-and-drop or upload an image to the Reference Image section of the Prompt panel. Then instruct the model on modifications you wish to make using active verbs. You may issue multiple instructions in the same prompt.

  • Change the king’s chalice from silver to gold. Give him a crown, and grow him a salt-and-pepper beard.
  • Change the image style to a scifi/fantasy vibe.
  • Use an anime style and give the noblemen and courtiers brightly-colored robes.

F2K editing supports multiple reference images, letting you transfer visual elements (subjects, style and background) from one to another. When prompting over multiple images, refer to them in order as “image 1,” “image 2,” and so forth.

  • Give the king in image 1 the crown that appears in image 2.
  • Transfer the style of image 1 to image 2.

Dealing with multiple reference images is tricky. There is no way to adjust the weightings of each image, and so you will have to be explicit in the prompt about which visual elements you are combining. If you cannot get the effect you are looking for by modifying the prompt, you may find success by changing the order of images.

Also be aware that each image significantly increases the model’s VRAM usage. If you run into memory errors, use a smaller (quantized) model, or reduce the number and size of the reference images.

Other Versions of F2K Available in the Model Manager

To find additional supported versions of F2K, type “FLUX.2” into the Starter Models search box. This will show you the following types of files:

  • FLUX.2 Klein 4B/9B (Diffusers) These are the full-size all-in-one diffusers versions of F2K which come bundled with the VAE and text encoder.
  • FLUX.2 Klein 4B/9B These are standalone versions of the full-size F2K which require installation of separate VAE and text encoders. Note that the 4B and 9B models require different text encoders, “FLUX.2 Klein Qwen3 4B Encoder” and “FLUX.2 Klein Qwen3 8B Encoder” respectively. (Not a misprint: use the 9B F2K model with the 8B text encoder!)
  • FLUX.2 Klein 4B/9B (FP8) These are the standalone versions quantized to 8 bits. The 4B model will run comfortably on macines with 8GB VRAM, while the 9B model will run on machines with 12GB or higher. As with all quantized versions, there is minor loss of generation accuracy.
  • FLUX.2 Klein 4B/9B (Q4) These are standalone versions that have been quantized to 4 bits, resulting in very small and fast models that can run on cards with 6-8 GB VRAM.

There is only one F2K VAE, and it happens to be same as the one used by FLUX.1 and Z-Image Turbo. However, there are several text encoder options:

  • FLUX.2 Klein Qwen3 4B Encoder Use this encoder with the F2K 4B versions. It also works with Z-Image Turbo.
  • Z-Image Qwen3 Text Encoder (quantized) This is a Q6-quantized version of the text encoder, that works with both F2K and ZiT. You may use this on smaller memory systems to reduce swapping of models in and out of VRAM.
  • FLUX.2 Klein Qwen3 8B Encoder Use this encoder with the F2K 9B versions. It is not compatible with ZiT.

You will find additional F2K models on HuggingFace and other model repositories, including the base models intended for fine-tuning and LoRA training. We have not exhaustively tested InvokeAI compatibility with all the available variants. Please report any incompatible models to InvokeAI Issues.

Many credits to @Pfannkuchensack for contributing F2K support.

Other Features in this Release

The other features in this release include:

Z-Image Turbo Variance Enhancer

ZiT tends to produce very similar images for a given prompt. To increase image diversity, @Pfannkuchensack contributed a Seed Variance Enhancer node which adds calibrated amounts of noise to the prompt conditioning prior to generation. You will find this feature in the Generation panel under Advanced Options. When activated, you will see two sliders, one for Variance Strength and the other for Randomize Percent. The first slider controls how much noise will be added to the conditioned prompt, and the second controls what proportion of the conditioning’s weights will be altered. Using the default randomization of 50% of the values, a variance strength of 0.1 will produce subtle variations, while a strength of 0.5 will produce very marked deviation from the prompt. Increasing the percentage of weights modified will also increase the level of variation.

Improved Support for High-Resolution FLUX.1 Images

A new denoising tuning algorithm, introduced by @Pfannkuchensack, increases the accuracy of FLUX.1 generations at high resolutions. When a FLUX.1 model is selected, a new DyPE option will appear in the Generation panel. Its settings are Off (the default) to disable the algorithm, Auto to automatically activate DyPE when rendering images greater than 1536 pixels in either dimension, and 4K Optimized to activate the algorithm with parameters that are tuned for 4K images. Note that if you do not have sufficient VRAM to generate 4K images, this feature will not help you generate them. Instead, generate a smaller image and use Invoke’s Upscaling feature.

Canvas high level transform smoothing

Another improvement contributed by @DustyShoe: The Canvas raster layer transform operation now supports multiple types of smoothing, thereby reducing the number of artifacts when an area is upscaled.

Text Search and Highlighting in the Image Metadata Tab

The Image Viewer’s info (🛈) tab now has a search field that allows you to rapidly search and highlight text in image metadata, details, workflow and generation graph. In addition, the left margin of the metadata display has been widened to make the display more readable.

Thanks to @DustyShoe for this improvement.

Bugfixes

Several bugs were caught and fixed in this release and are listed in the detailed changelog below. Thanks to first-time contributors @kyhavlov and @aleyan for the bugs they caught and fixed.

Installing and Updating

The Invoke Launcher is the recommended way to install, update and run Invoke. It takes care of a lot of details for you - like installing the right version of python - and runs Invoke as a desktop application.

Note: With recent updates to torch, users on older GPUs (20xx and 10xx series) will likely run into issues with installing/updating. We are still evaluating how we can support older GPUs, but in the meantime users have found success manually downgrading torch. Head over to discord if you need help.

Follow the Quick Start guide to get started with the launcher.

If you don’t want to use the launcher, or need a headless install, you can follow the manual install guide.

Translation Credits

Many thanks to the following language translators who contributed to this release: @Harvester62 (Italian) and @DustyShoe (Russian).

Also many thanks to Weblate for granting InvokeAI a free Open Source subscription to use its translation management service.

What’s Changed

New Contributors

Full Changelog: https://github.com/invoke-ai/InvokeAI/compare/v6.10.0…v6.11.0

v6.10.0

InvokeAI v6.10.0

This is the first InvokeAI Community Edition release since the closure of the commercial venture, and we think you will be pleased with the new features and capabilities. This release introduces backend support for the state-of-the-art Z-Image Turbo image generation models, and multiple frontend improvements that make working with InvokeAI an even smoother and more pleasurable experience.

The Z-Image Turbo Model Family

Z-Image Turbo (ZiT) is a bilingual image generation model that manages to combine high performance with a small footprint and excellent image generation quality. It excels in photorealistic image generation, renders both English and Chinese text accurately, and is easy to steer. The full model will run easily on consumer hardware with 16 GB VRAM, and while quantized versions will run on significantly smaller cards, with some loss of precision.

With this release InvokeAI runs almost all released versions of ZiT, including diffusers, safetensors, GGUF, FP8 and quantified versions. However, be aware that the FP8 scaled weights models are not yet fully supported and will produce image artifacts. In addition, InvokeAI supports text2image, image2image, ZiT LoRA models, controlnet models, canvas functions and regional guidance. Image Prompts (IP) are not supported by ZiT, but similar functionality is expected when Z-Image Edit is publicly released.

To get started using ZiT, go to the Models tab and from the Launchpad select the Z-Image Turbo bundle to install all the available ZiT related models and dependencies (rougly 35 GB in total). Alternatively, you can select individual models from the Starter Models tab, and search for “Z-Image.” The full and Q8 models will run on a 16 GB card. For cards with 6-8 GB of VRAM, choose the smaller quantized model, Z-Image Turbo GGUF Q4_K. Note that when using one of the quantized models, you will also need to install the standalone Qwen3 encoder and one of the Flux VAE models. This will be handled for you when you install a ZiT starter model.

When generating with these models it is recommended to use 8-9 steps and a CFG of 1. Be aware that due to ZiTs strong prompt following it does not generate as much image diversity as other models you may be used to. One way to increase image diversity is to create a custom workflow that adds noise to the Z-Image Text Encoder using @Pfannkuchensack’s Image Seed Variance Enhancer Node.

In addition to the default Euler scheduler for ZiT we offer the more accurate but slower Heun scheduler, and a faster but less accurate LCM scheduler. Note that the LCM and Heun schedulers are still experimental, and may not produce optimal results in some workflows.

A big shout out to @Pfannkuchensack for his critical contributions to this effort.

New Workflow Features

We have two new improvements to the Workflow Editor:

  • Workflow Tags: It is now possible to add multiple arbitrary text tags to your workflows. To set a tag on the current workflow, go to Details and and scroll down to Tags. Enter a comma-delimited of tags that describe your workflow, such as “image, bounding box”, and save. The next time you browse your workflows, you will see a series of checkboxes for all the unique tags in your workflow connection. Select the tag checkboxes individually or in combination to filter the workflows that are displayed. This feature was contributed by @Pfannkuchensack.
  • Prompt Template Node: Another @Pfannkuchensack workflow contribution is a new Prompt Template node, which allows you to apply any of the built-in or custom prompt style templates to a prompt before passing it onward to generation.

Prompt Weighting Hotkeys

@joshistoast has added a neat feature for adjusting the weighting of words and phrases in the prompt. Simply select a word or phrase in the prompt textbox and press Ctrl-Up Arrow to increase the weight of the selection (by adding ”+” marks) or Ctrl-Down Arrow to decrease the weighting.

Limitations: The prompt weighting does not work properly with numeric weights, nor with prompts that contain the .add() or .blend() functions. This will be fixed in the next point release.

Hotkey Editor

Speaking of hotkeys, @Pfannkuchensack and @joshistoast contributed a new user interface for editing hotkeys. Any of the major UI functions, such as kicking off a generation, opening or closing panels, selecting tools in the canvas, gallery navigation, and so forth, can now be assigned a key shortcut combination. You can also assign multiple hotkeys to the same function.

To access the hotkey editor, go to the Settings (gear) menu in the bottom left, and select Hotkeys.

Bulk Operations in the Model Manager

You can now select multiple models in the Model Manager tab and apply bulk operations to them. Currently the only supported operation is to Delete unwanted models, but this feature will be expanded in the future to allow for model exporting, archiving, and other functionality.

This feature was contributed by @joshistoast, based on earlier work by @Pfannkuchensack .

Masked Area Extraction in the Canvas

It is now possible to extract an arbitrary portion of all visible raster layers that are covered by the Inpaint Mask. The extracted region is composited and added as a new raster layer. This allows for greater flexibility in the generation and manipulation of raster layers.

Thanks to @DustyShoe for this work.

PBR Maps

@blessedcoolant added support for PBR maps, a set of three texture images that can be used in 3D graphics applications to define a material’s physical properties, such as glossiness. To generate the PBR maps, simply right click on any image in the viewer or gallery, and select “Filters -> PBR Maps”. This will generate PBR Normal, Displacement, and Roughness map images suitable for use with a separate 3D rendering package.

New FLUX Model Schedulers

We’ve also added new schedulers for FLUX models (both dev and schnell). In addition to the default Euler scheduler, you can select the more accurate but slow Heun scheduler, and the faster but less accurate LCM scheduler. Look for the selection under “Advanced Options” in the Text2Image settings panel, or in the FLUX Denoise node in the workflow editor. Note that the LCM and Heun schedulers are still experimental, and may not produce optimal results in some workflows.

Thanks to @Pfannkuchensack for this contribution.

SDXL Color Compensation

When performing SDXL image2image operations, the color palette changes subtly and the discrepancy becomes increasingly obvious after several such operations. @dunkeroni has contributed a new advanced option to compensate for this color drift when generating with SDXL models.

Option to Release VRAM When Idle

InvokeAI tends to grab as much GPU VRAM as it needs and then hold on to it until the model cache is manually cleared or the server is restarted. This can cause an annoyance for people who need the VRAM for other tasks. @lstein added a new feature that will automatically clear the InvokeAI model cache and release its VRAM after a set period of idleness. To activate this feature, add the configuration option model_cache_keep_alive_min to the invokeai.yaml configuration file. It takes a floating point number corresponding to the number of minutes of idleness before VRAM is released. For example, to release after 5 minutes of idleness, enter:

model_cache_keep_alive_min: 5.0

Setting this value to 0 disables the feature. This is also the default if the configuration option is absent.

Bugfixes

Multiple bugs were caught and fixed in this release and are listed in the detailed changelog below.

Installing and Updating

The Invoke Launcher is the recommended way to install, update and run Invoke. It takes care of a lot of details for you - like installing the right version of python - and runs Invoke as a desktop application.

Note: With recent updates to torch, users on older GPUs (20xx and 10xx series) will likely run into issues with installing/updating. We are still evaluating how we can support older GPUs, but in the meantime users have found success manually downgrading torch. Head over to discord if you need help.

Follow the Quick Start guide to get started with the launcher.

If you don’t want to use the launcher, or need a headless install, you can follow the manual install guide.

New Contributors

Translation Credits

Many thanks to Riccardo Giovanetti (Italian) and RyoKoba (Japanese) who contributed their time and effort to providing translations of InvokeAI’s text.

What’s Changed

Full Changelog: https://github.com/invoke-ai/InvokeAI/compare/v6.9.0…v6.10.0

v6.9.0

This release focuses on improvements to Invoke’s Model Manager. The changes are mostly internal, with one significant user-facing change and a data migration.

On first run after installing this release, Invoke will do some data migrations:

  • Run-of-the mill database updates.
  • Update some model records to work with internal Model Manager changes, described below.
  • Restructure the Invoke-managed models directory into a flat directory structure, where each model gets its own folder named by the model’s UUID. Models outside the Invoke-managed models directory are not moved.

If you see any errors or run into any problems, please create a GH issue or ask for help in the #new-release-discussion channel of the Invoke discord.

Model Installation Improvements

Invoke analyzes models during install to attempt to identify them, recording their attributes in the database. This includes the type of model, its base architecture, its file format, and so on. This release includes a number of improvements to that process, both user-facing and internal.

Unknown Models

Previously, when this identification fails, we gave up on that model. If you had downloaded the model via Invoke, we would delete the downloaded file.

As of this release, if we cannot identify a model, we will install it as an Unknown model. If you know what kind of model it is, you can try editing the model via the Model Manager UI to set its type, base, format, and so on. Invoke may be able to run the model after this.

If the model still doesn’t work, please create a GH issue linking to the model so we can improve model support. The internal changes in this release are focused on making it easier for contributors to support new models.

Invoke-managed Models Directory

Previously, as a relic of times long past, Invoke’s internal model storage was organized in nested folders: <models_dir>/<type>/<base>/model.safetensors. Many moons ago, we didn’t have a database, and models were identified by putting them into the right folder. This has not been the case for a long time.

As of this release, Invoke’s internal model storage has a normalized, flat directory structure. Each model gets its own folder, named its unique key: <models_dir>/<model_key_uuid>/model.safetensors.

On first startup of this release, Invoke will move model files into the new flat structure. Your non-Invoke-managed models (i.e. models outside the Invoke-managed models directory) won’t be touched.

We understand this change may seem user-unfriendly at first, but there are good reasons for it:

  • This structure eliminates the possibility of model name conflicts, which have caused of numerous hard-to-fix bugs and errors.
  • It reinforces that the internal models directory is Invoke-managed:
    • Adding models to this directory manually does not add them to Invoke. With the previous structure, users often dropped models into a folder and expected them to work.
    • Deleting models from this directory or moving them in the directory causes the database to lose track of the models.
  • It obviates the need to move models around when changing their type and base.

Refactored Model Identification system

Several months ago, we started working on a new API to improve model identification (aka “probing” or “classification”). This process involves analyzing model files to determine what kind of model it is.

As of this release, the new API is complete and all legacy model identification logic has been ported to it. Along with the changes in #8577, the process of adding new models to Invoke is much simpler.

Model Identification Test Suite

Besides the business logic improvements, model identification is now fully testable!

When we find a model that is not identified correctly, we can add that model to the test suite, which currently has test cases for 70 models.

Models can be many GB in size and are thus not particularly well-suited to be stored in a git repo. We can work around this by creating lightweight representations of models. Model identification typically relies on analyzing model config files or state dict keys and shapes, but we do not need the tensors themselves for this process.

This allows us to strip out the weights from model files, leaving only the model’s “skeleton” as a test case. The 70-model test suite is currently about 115MB but represents hundreds of GB of models.

Installing and Updating

The Invoke Launcher is the recommended way to install, update and run Invoke. It takes care of a lot of details for you - like installing the right version of python - and runs Invoke as a desktop application.

Note: With recent updates to torch, users on older GPUs (20xx and 10xx series) will likely run into issues with installing/updating. We are still evaluating how we can support older GPUs, but in the meantime users have found success manually downgrading torch. Head over to discord if you need help.

Follow the Quick Start guide to get started with the launcher.

If you don’t want to use the launcher, or need a headless install, you can follow the manual install guide.

What’s Changed

Full Changelog: https://github.com/invoke-ai/InvokeAI/compare/v6.8.1…v6.9.0

v6.8.1

This patch release fixes the Exception in ASGI application startup error that prevents Invoke from starting.

The error was introduced by an upstream dependency (fastapi). We’ve pinned the fastapi dependency to the last known working version.

Installing and Updating

The Invoke Launcher is the recommended way to install, update and run Invoke. It takes care of a lot of details for you - like installing the right version of python - and runs Invoke as a desktop application.

Note: With recent updates to torch, users on older GPUs (20xx and 10xx series) will likely run into issues with installing/updating. We are still evaluating how we can support older GPUs, but in the meantime users have found success manually downgrading torch. Head over to discord if you need help.

Follow the Quick Start guide to get started with the launcher.

If you don’t want to use the launcher, or need a headless install, you can follow the manual install guide.

What’s Changed

Full Changelog: https://github.com/invoke-ai/InvokeAI/compare/v6.8.0…v6.8.1

v6.8.0

This minor release includes a handful of fixes and enhancements.

Fixes

  • When accepting raster layer adjustments, the opacity of the layer was “baked” in.
  • Corrected help text for non-in-place model installation. Previously, the help text said that a non-in-place model install would copy the model files. This is incorrect; it moves them into the Invoke-managed models dir.
  • Failure to queue generations with an error like Failed to Queue Batch / Unknown Error.

Enhancements

  • Added a crop tool. For now, it is only enabled for Global Ref Images.
    • Click the crop icon on the Ref Image preview to open the tool.
    • Adjust the crop box and click apply to save the cropped image for that ref image.
    • To revert, open the crop tool, click Reset, then Apply to revert to the original image.
    • We’ll explore integrating this new tool elsewhere in the app in a future update.
  • Improved Model Manager tab UI. Thanks @joshistoast!
  • Keyboard shortcuts to navigate prompt history. Use alt/option+up/down to move through history.
  • Support for the NOOB-IPA-MARK1 IP Adapter. Thanks @Iq1pl!

Internal

  • Support for dynamic model drop-downs in Workflow Editor. This change greatly reduces the amount of frontend code changes needed to support a new model type. Node authors may need to update their nodes to prevent warnings from being displayed. However, there are no breakages expected. See #8577 for more details.

Installing and Updating

The Invoke Launcher is the recommended way to install, update and run Invoke. It takes care of a lot of details for you - like installing the right version of python - and runs Invoke as a desktop application.

Note: With recent updates to torch, users on older GPUs (20xx and 10xx series) will likely run into issues with installing/updating. We are still evaluating how we can support older GPUs, but in the meantime users have found success manually downgrading torch. Head over to discord if you need help.

Follow the Quick Start guide to get started with the launcher.

If you don’t want to use the launcher, or need a headless install, you can follow the manual install guide.

What’s Changed

New Contributors

Full Changelog: https://github.com/invoke-ai/InvokeAI/compare/v6.7.0…v6.8.0

v6.7.0

This minor release includes improved object selection on Canvas, layer adjustments, prompt history and a handful of other enhancements and fixes.

Select Object v2

We’ve made some major improvements to object selection.

  • Segment Anything v2 is now supported. You can choose between SAM1 and SAM2. We’ve found that SAM2 is much faster than SAM1, but often does not perform as well, so we left SAM1 as an option.
  • You may now draw a box around the target object. The box doesn’t need to be exact - sometimes, you can get better results by making it a bit smaller than the target object. Points are still supported and can be used independently or as a refinement for a box.
  • Holding shift while clicking creates an exclude point if you have include selected. If you have selected exclude, holding shift will instead create an include point.
  • You can now provide a text prompt instead of a box and points. Use very simple language for best results. Internally, this uses Grounding DINO to identify the target.

Raster Layer Adjustments

Right click a Raster Layer to add adjustments. Adjustments are non-destructive, though you can accept them to bake them into the layer.

You can adjust brightness, contrast, saturation, temperature, tint, and sharpness, or use the curves editor to adjust each channel independently.

Thanks @dunkeroni for implementing this very useful feature.

Prompt History

There’s a new button in the Positive Prompt box for prompt history. Your last 100 unique prompts are stored for easy recall. You can search them, delete individual prompts, or clear the whole list.

Enhancements

  • Improved object selection on Canvas.
  • Raster layer adjustments. Thanks @dunkeroni!
  • Support for mathematical expressions in number input fields. Currently, these are only enabled for fields in the Workflow Editor (including Builder Forms). Thanks @csehatt741!
  • Prompt history for Positive Prompt.
  • Queue list now sorts with newest on top. You can reverse the sort if you want, to restore the previous sorting. Thanks @csehatt741!
  • Updated translations. Thanks @Harvester62 @Linos1391!

Fixes

  • Fixed an issue that prevented you from using LoRA weights outside the range -1 to 2.
  • Fixed an issue where LoRA settings could be lost on refresh.
  • Fixed an issue where LoRAs with weights outside the range -1 to 2 were not able to be recalled from metadata.
  • Fixed an issue where popovers like the Canvas Settings popover were obscured by other UI elements.
  • Fixed a path traversal vulnerability affecting the bulk downloads API.

Installing and Updating

The Invoke Launcher is the recommended way to install, update and run Invoke. It takes care of a lot of details for you - like installing the right version of python - and runs Invoke as a desktop application.

Note: With recent updates to torch, users on older GPUs (20xx and 10xx series) will likely run into issues with installing/updating. We are still evaluating how we can support older GPUs, but in the meantime users have found success manually downgrading torch. Head over to discord if you need help.

Follow the Quick Start guide to get started with the launcher.

If you don’t want to use the launcher, or need a headless install, you can follow the manual install guide.

What’s Changed

Full Changelog: https://github.com/invoke-ai/InvokeAI/compare/v6.6.0…v6.7.0

v6.6.0

This is a minor release, adding a few QoL improvements and fixes.

Enhancements

  • Canvas Color Picker has foreground and background colors. Switch between them with x. Press d to reset them to black and white. Thanks @csehatt741!
  • You can set a default weight setting for LoRAs in the Model Manager. When you add the LoRA, it will start at the default weight. Thanks @csehatt741!
  • Canvas Brush/Eraser width renders an in-line slider when there is enough space instead of showing the slider in a popover.
  • Updated translations. Thanks @Harvester62!

Fixes

  • Always delete LoRAs when recalling all metadata. Thanks @csehatt741!
  • Incompatible LoRAs being enabled prevents you from clicking Invoke.
  • Fixed an issue where it was possible to drag a tab panel to another location in the UI on Chrome and Launcher (Firefox was unaffected).
  • Internal file organization fix for docker builds.
  • Fix an issue where progress images were super tiny (again).
  • Fix an issue where no fallback was rendered in the viewer when no image is selected.
  • Fix an issue where a single middle-mouse click on Canvas would activate the View tool (i.e. drag-to-pan), and you had to click again to deactivate it.
  • Fix an issue in the Viewer where the last-generated image would briefly show after the current generation finishes.

Installing and Updating

The Invoke Launcher is the recommended way to install, update and run Invoke. It takes care of a lot of details for you - like installing the right version of python - and runs Invoke as a desktop application.

Note: With recent updates to torch, users on older GPUs (20xx and 10xx series) will likely run into issues with installing/updating. We are still evaluating how we can support older GPUs, but in the meantime users have found success manually downgrading torch. Head over to discord if you need help.

Follow the Quick Start guide to get started with the launcher.

If you don’t want to use the launcher, or need a headless install, you can follow the manual install guide.

What’s Changed

Full Changelog: https://github.com/invoke-ai/InvokeAI/compare/v6.5.1…v6.6.0

This site was designed and developed by Aether Fox Studio.