Last week, Meta unveiled a new AI image generation tool called Imagine, which uses the company’s AI model, Llama 3, to produce high-quality images. The goal of Imagine is to help users let their imaginations run wild with assistance from the company’s AI model. Users can generate images in real time by typing a prompt that describes what they want to create.
Imagine is currently being tested in beta on WhatsApp and the Meta AI website, where it allows images to evolve with each typed letter. The feature relies on Meta’s latest open-source large language model, Llama 3, which is designed to improve image quality and incorporate text into images more effectively. Once an image is generated, users can animate it or transform it into a GIF or apply new styles.
In addition to Imagine, Meta introduced several other updates to its AI assistant, such as providing tailored recommendations for dining or interior design inspiration. However, despite these advancements, Meta is still working hard to catch up with competitors like OpenAI and Anthropic. The company continues to refine its larger models of Llama 3 to improve their capabilities. Meta’s CEO Mark Zuckerberg has described the developments as “pretty wild,” with future models expected to rival those released by other AI companies earlier this year.