We’ve been working hard on Logo Diffusion V.3 since November, and we’re finally ready to release it! with a new logo design A.I model built from scratch, this is our most ambitious release so far and below we’ll detail the improvements and new features we’ve added to your Logo Diffusion experience.

We’ve been working hard on Logo Diffusion V.3 since November, and we’re finally ready for its release! With a new logo design A.I model built from scratch, this is our most ambitious release to date.

Below you can find all the details of the improvements and the new features we’ve added to your Logo Diffusion experience.

We’ve also prepared a quick video (below) to go over some of the new features in V.3.

V.3 New Logo Design Model:

We decided to train our model from scratch with an emphasis on prompt accuracy, so now you can get the results you want faster and without burning through too many credits in the process.

Below you can see a few examples of the same prompt ran in Logo Diffusion V.2 vs. Logo Diffusion V.3:

Seeing the examples above, V.3 shows much better prompt understanding as well as accuracy in both the “Mascot” and “Cartoon” examples. This improvement carries over to the ability to generate text logos which can be seen in the Logotype example, where V.2 understood the prompt correctly only 1 out of 4 results, while V.3 got it right 3 out of 4 times.

You can try the new model for yourself on Logo Diffusion for free!

17 new New Styles (45 Total) :

We improved our existing 28 styles and added 17 more to bring our total to 45 Art Styles to choose from, all of which will be running on our new and improved model for better quality and prompt accuracy. They can also be used with our image-to-image workflow, so you can easily change the style of your logos; any logo or design can be transformed into any of the 45 Art Styles in just a few clicks.

New Dashboard and workspaces:

- Discover and Remix:

Logo Diffusion V.3 comes with an updated dashboard that features a Discover and Remix section that gets updated weekly, where you can find new ideas and workflows to get inspired or use them as a basis for your next design.

We improved our existing 28 styles and added 17 more to bring our total to 45 Art Styles to choose from, all of which will be running on our new and improved model for better quality and prompt accuracy. They can also be used with our image-to-image workflow, so you can easily change the style of your logos; any logo or design can be transformed into any of the 45 Art Styles in just a few clicks.

The Dashboard also features more workspaces to help you work faster and more efficiently. These workspaces can be accessed from our new Dashboard page or the sidebar, we’ll go over them one by one below.

- Sketch to Image workspace:

We built this workspace specifically for the sketch-to-image and image-to-image workflows to help you work faster and compare results more easily, but of course, you will still have the option to use our main text-to-image workflow.

- Image Upscaler Workspace:

Upscaling images is not only limited to designs you generate using Logo Diffusion, you can now upscale any image you already have using our new Upscaling Workspace.

- Background Removal Workspace:

We did the same for background removal, so now you can remove the background from any image you upload, then download it or send it to our Image-to-Image workspace to keep working on it.

- Image to Vector Workspace:

Our Vectorization tool is no different, you can also turn any image you have into an editable vector .SVG file.

UI improvements:

- Improved Color selection UI:

Color selection has become more streamlined and easier to use, with a global setting that works on all styles, and the ability to switch between solid and gradient colors (more color options coming soon).

- Easy Access to Negative prompt:

The negative prompt box can now be easily accessed in one click via the “Exclude” button.

- Style-Specific settings:

The 9 main logo styles, as well as the “3D text” style, now all come with custom style settings. You can select different parameters for each style to help you get more specific results, or you can set it to “Randomize” to get new versions of your prompt every time you click generate.

And that’s pretty much it for our logo Diffusion V.3 release, We cant wait to see what you’ll come up with using it.

You can try the new model for yourself on Logo Diffusion for free!

You can reach out to us at Discord, Facebook Groups, Twitter, Instagram, and Tiktok.