Category: Images

Imagine being able to say “take the person on the left in image 1 and the middle person in image 2 and have them [whatever you went them to do, wherever you want them to do it]. Or telling it to deblur an image, or add or remove things when combining multiple images or parts of images.

Omnigen can do this and much more. You just tell it what you want it to do to the image and it does it. You can see it in action at https://www.youtube.com/watch?v=PCL9SAlHqzw

And try it out at https://huggingface.co/spaces/Shitao/OmniGen

Warning: Being designed by geeks, it’s not the most intuitive, and it can cost you in credits after a while. If you have a powerful enough PC and graphics card, you can install it locally and use it for free with no limits.

OpenAI, the company who makes ChatGPT just raised the bar in image generation. And within 24 hours my favorite image generator for the last year, Ideogram 2.0, released version 3.0. Both of these are game changers.

OpenAI played it smart this time, unlike when they announced their quantum leap in video generation, Sora, in February of 2024 and then didn’t release it until December, by which time many other companies had gone beyond it. So its release was kind of like a damp squib.

This time they said nothing and just released it and became the best image generator there is. They have finally retired my least favorite generator DALL-E3, thank goodness. They’re just calling it 4o Image Generation and you can access it by typing “Create Image” (or clicking on the 3 View Tools dots) right within ChatGPT, even the free version (which has limits as to how many images you can create).

What’s special about it? You can refine and edit your image via natural conversations, keeping what you like exactly as it is and only changing what you want changed, it delivers exactly what you ask it to, its handling of text, even long text is phenomenal, it’s photorealistic or whatever you want it to be, you can show it an image as a guide, and more. It can create beautiful promo pieces fully. It can accept a single image of you, or anybody, and you can place you or them in any environment. It will maintain character consistency from image to image, and more.

It uses a similar technology to the Google AI Studio one that I wrote about on Monday, but in my tests, It is better.

You can find out more at https://openai.com/index/introducing-4o-image-generation/

And yesterday Ideogram 3.0 came out, which is also a wonderfully enhanced version, with a lot of similar features. To prevent this post from becoming a book, you can see and read more about it at https://about.ideogram.ai/3.0

And now there’s also Reve, which is also excellent!  https://preview.reve.art/app/explore

Flux Kontext is a new AI tool that lets you edit images using simple text instructions. Unlike other tools that require complex prompts or manual editing, Flux Kontext allows you to make specific changes—like altering colors, modifying text, or adjusting objects, while keeping the rest of the image intact. This makes it easier for everyday users to tweak images without needing advanced skills. What sets Flux Kontext apart is its ability to understand both the image and your instructions, enabling precise edits and maintaining consistency across multiple changes. Whether you’re updating a product photo or creating a series of images with the same character, Flux Kontext helps you achieve your goals with less effort. You can try it out through platforms like Flux AI Studio and BFL AI.

Qwen Image Edit is a fresh take on image editing that lets you transform your photos simply by typing what you’d like done, no tricky software or jargon needed. Want to add a fun neon sign, remove a stray hair, or even rotate an object to another angle? Just say so, and the tool takes care of it. Plus, it handles both English and Chinese text seamlessly, preserving the original font and style which not many tools can do yet.Unlike traditional editors that redraw your whole image, Qwen can make broad creative changes (like a style shift) or very precise tweaks (like changing just one letter) while keeping the rest of your image untouched. It’s already earning praise for being open source, easy to try, and impressively thoughtful in how it edits. Learn more..

Nano Banana, also known as Gemini 2.5 Flash, is being hailed for something that sets it apart from many image tools: character consistency. While lots of AI generators can create stunning pictures, Nano Banana makes it possible to edit an image multiple times—changing hairstyles, clothes, or even the background, while still keeping the subject recognizable. This is especially useful for creators who want to tell a story or keep a look steady across different edits, or even videos.

What makes it even more unique is its versatility. Beyond simple edits, it can handle transformations like turning a 2D photo into 3D, restoring old pictures, or zooming in and out to reveal new perspectives. Other tools may focus on one specialty, but Nano Banana brings a wide range of editing options together in one place, making it practical for everything from product mockups to personal photo experiments. Currently, you can access Nano Banana through multiple platforms including Google AI Studio for full access (with free, watermarked outputs), Google Gemini for lighter use with some limitations, and LM Arena, which offers watermark-free access.

Seedream 4.0 from ByteDance tries to simplify how you create and edit images. Now, instead of switching between tools, you can start with a blank canvas or upload your own picture, then use a text prompt to adjust or change things, remove someone, change the setting, tweak lighting, etc. while keeping everything else the same. It also accepts multiple reference images, so you can mix styles or pull ideas from several visuals at once. 

What makes Seedream 4.0 stand out is its blend of quality and flexibility. It supports high-resolution output (up to 4K), works faster than earlier versions, and handles “knowledge-driven” edits like illustrations or diagrams more reliably. It claims to be better than Nano Banana for image editing, but they are both excellent, in my experience.