While it might seem like we’ve been given more than enough AI image generators, for better or worse, new models are popping up all the time. Last week we say Elon Musk‘s X roll out a chaotic new version of Grok AI, and now Google has unveiled yet another image model that it claims is even more effective than its rivals.
Google claims that Imagen 3 is capable of “generating images with even better detail, richer lighting and fewer distracting artifacts than our previous models,” (and we were already impressed with with Imagen 2. And it all boils down the model’s much more comprehensive understanding of text prompts.
“We’ve significantly improved Imagen 3’s ability to understand prompts, which helps the models generate a wide range of visual styles and capture small details from longer prompts,” Google announces. “We’ve designed Imagen 3 to generate high-quality images in a wide range of formats and styles, from photorealistic landscapes to richly textured oil paintings or whimsical claymation scenes. Imagen 3 also understands prompts written in natural, everyday language, making it easier to get the output you want without complex prompt engineering.”
Imagen 3 is now available to try for free right now for those in the US. Through its ImageFX platform, users can try out their own detailed text prompts to generate images.
Google has also released a detailed paper on Imagen 3, detailing how the model’s output “is preferred over other state-of-the-art models at the time of evaluation.” In other words, Google thinks it’s better than its competitors, some of which are paid. And judging by the response on Twitter (sorry, X), it seems general users agree.
Google’s Imagen 3 via ImageFX has generated some of the most realistic AI portrait images I’ve seen so far 📸 pic.twitter.com/e3kkgRd74hAugust 19, 2024
But while Google’s new offering might be impressive, there’s certainly a feeling that the tide might be (somewhat) turning against AI. From Procreate’s new fiercely anti-AI stance to the backlash against the chaos of Grok, not everyone is celebrating the advent of ever more realistic generative models.