Google has quietly launched the Imagen 3 AI image generator. The latest version of the search titan’s text-to-image AI model is now available to everyone residing in the US. Notably, the company introduced the Imagen 3 AI image generator during the I/O event in May this year. However, the tech giant seems to have made it available via its Vertex AI platform in the past several days.
Google’s latest AI image generator claims to offer “better detail and richer lighting”
According to Google, the Imagen 3 is its highest-quality text-to-image artificial intelligence model. The Google AI image generator is touted to be capable of “generating images with even better detail, richer lighting and fewer distracting artifacts than the previous models”. The company has significantly improved the AI generator’s ability to understand text prompts.
The new Imagen 3 model is capable of generating a wide range of visual styles. Furthermore, the latest Google AI image generator can capture small details from the longer user text prompts. Besides, the Imagen 3 will be offered in multiple variants. Google says that all of the versions are optimized for different types of tasks, “from generating quick sketches to high-resolution images”.
“To help Imagen 3 capture nuances like specific camera angles or compositions in long, complex prompts, we added richer detail to the caption of each image in its training data”, Google further explained about the new abilities of the platform.
The Imagen 3 AI model can be accessed through Google’s AI Test Kitchen. It is available for the users of the ImageFX and Vertex AI platforms in the United States. The company is expected to expand the availability of the new tool in other countries in the near future.
Imagen 3’s launch coincides with the release of xAI’s Grok-2 image generator
Google’s Imagen 3 AI image creation platform’s launch coincides with the release of Elon Musk’s xAI’s Grok-2 AI image generator. Notably, Musk’s AI image creator doesn’t appear to have any safety filters. Apparently, Grok has been creating all sorts of reckless photos recently, which many users have shared on X.
The images generated by the Grok-2 platform are pretty realistic, but it doesn’t have any filters. It can be used to generate something potentially brazen like the former US president Barack Obama stabbing Joe Biden.
Google’s AI image creation tools have run into some problems as well in the past. However, Google says the Imagen 3 was “built with our latest safety and responsibility innovations, from data and model development in production”. The company claims to have used extensive filtering and data labeling to minimize the creation of any harmful content.