Krea 3

“some of the biggest open-weight bangers (Krea 2, Ideogram 4, and the kind of slept-on Boogu-Image 0.1)”

“You can get results a billion times better than the horrible, disgusting look of ChatGPT in less than a couple of seconds”

“The few generations I did looked promising but their whole vibe is just sell sell and feels so cheap.”

“Not interesting unless completely open and locally sourcable.”

“Incredibly aggressive dark patterns in this sign up flow just to try the model.”

“this exact artifact appeared in Qwen, Grok, and now Krea”

“is that really applicable to a mistake like the text on the T-shirt extending over the suit jacket”

“the trial period the email you send out for new signups seems to be broken”

“None of these advances matter if the hyper-Puritan guardrails continue to be in place.”

“Not only open weight but even better terms in the license!”

“self-hosted capable with enterprise license agreement, contact us...”

“Krea2 was awesome because it was open”

“If I used Krea, it would be because I fine tuned/built a workflow on top of it to do something GPT image couldn't/wouldn't do.”

“it still should be commended for its ability to render small text details”

“The signage text on the fuzzy characters scene was scrambled.”

“No one is forcing you to employ this specific dark pattern regardless of how you claim your competition behaves.”

“Krea 2 ranks among the highest of the local models”

“Sneaky pricing. Offers you plans when you sign up, if you decline them, it offers you a discount.”

“the team is pretty happy with the text to image of Krea 2”

“the frontier-level capabilities and new capabilities should be available for people to get a great experience”

“scored pretty well in my GenAI benchmark”

“has a community license [2] that is fairly permissive”

“Krea can usually get close enough to a reference that it's not a big deal”

“among the best right now for text2image”

“I've even got it working locally on my M5 iPad Pro.”

“Krea2 has been remarkable for me so far”

“If you can get around the restrictive license, output speed, and JSON prompting”

“"alignment training" is pretty much a code word for censorship/lobotomy”

“It really punched above its weight class, scoring the highest among locally hostable models”

“it literally takes only a few minutes on Krea; definitely not "N hours for GPUs to churn"”

“I like the T2I from what I've seen, mind. Perhaps more than Images 2.0 or even NB2.”

“It's pretty great that you are providing the undistilled model on day 0.”

“we find internally that this model is used more than Nano Banana for many cases like moodboarding”

“Krea 2 produces the same blurry, airbrushed output as qwen image”

“We are on par with Nano Banana in terms of image quality”

“This model is actually meant to be hackable/fine-tunable”

“Definitely encourage you to test the models. We tried to optimize for realistic focus and not over-sharpening”

“if you want to test it out and push realism”

“It's a good model sadly the use of the qwen vae is a bit of a downer.”

“Modern tools like Krea or Magnific are no longer just "enter prompt, get image"”