
Google has released its Imagen 3 AI-image generator to all Gemini users. It allows users to create photos from text prompts using Generative AI. The feature is now available for all Gemini users, including free and paid subscribers. Here are the details.
Imagen 3 Now Available in Gemini for Free
Imagen 3 is the most advanced AI image generator by Google. It creates high-quality images with better details, richer lighting, and fewer distracting artefacts. It is powered by Gemini’s top-end large language models, enabling better text rendering and understanding of text prompts.
The feature is directly available in the Gemini app and website. Users can ask the chatbot to generate any image with a text prompt. Gemini will use Imagen 3 to create a relevant image. The photos have a 1:1 aspect ratio with a resolution of 2048×2048.
If the result is unsatisfactory, users can ask Gemini to make changes to the photo using additional text prompts. Note that Imagen 3 and Gemini are designed to maintain privacy and security, and will not generate images of notable people, celebrities, adult or immature content, etc. If your text prompt doesn’t sit well with Google’s safety guidelines, Gemini will refuse to create an image.
Here are some samples of images generated by Imagen 3 through Gemini. We also added captions with the text prompt that we used:
Our experience with Imagen 3 in Gemini was fascinating. The chatbot was fast and generated an image in less than 10 seconds. We found it to be faster than Meta’s AI image generator.
Unlike Meta, Imagen 3 in Gemini does not add any watermark to the image. Hence, it becomes difficult to identify whether the photo was AI-generated. This is both, an advantage and a disadvantage, depending on the situation.
To recall, Gemini already could generate images at the time of its launch. However, the chatbot faced huge backlash as it responded with highly irrelevant images, with poor accuracy. Google quickly acknowledged the issue and disabled the image generation in Gemini in February 2024. The feature has finally made a comeback today, in the form of Imagen 3.