(Credit: Google)
Google has opened up access to its Imagen 3 text-to-image generator.
After launching for select users in private preview at Google I/O, Imagen 3 is now available to all US users via ImageFX, VentureBeat reports.
Google describes Imagen 3 as its "highest quality text-to-image model yet." The tool "generates an incredible level of detail, producing photorealistic, lifelike images, with far fewer distracting visual artifacts than our prior models."
Google promises better understanding of natural language, meaning it picks up on the "intent behind your prompt and incorporates small details from longer prompts." It also excels at rendering text on images, which "opens up possibilities for generating personalized birthday messages, title slides in presentations and more."
As VentureBeat notes, however, some early testers were frustrated by "the censorship filter which is far more sensitive to benign words."
"It's pretty good but I'm working harder with higher error results," one of them wrote on Reddit. "I still think Google have the best image generation out of all the ones I've used but this Imagen 3 is a bit of a regression for me personally" versus Imagen 2.
For a more in-depth look at Imagen 3, the Google team published a research paper about the tech.
Google may be trying to avoid what happened with Gemini image generation earlier this year. After users complained about Gemini creating historical images that depicted people of different races in photos where they likely would have not been present, Google paused image generation of people completely.
Elon Musk apparently doesn't have similar concerns. Earlier this week, his xAI startup launched Grok-2, which includes an image generator that has few, if any, guardrails.