Amazon Web Services unveils Nova, a new suite of multimodal AI models

Amazon Web Services (AWS) introduced a new family of AI models called Nova at its re:Invent conference. The Nova family includes four text-focused models: Micro, Lite, Pro, and Premier, with the first three available now and Premier set to launch in early 2025. In addition to text models, AWS launched two generative media models: Nova Canas for image generation and Nova Reel for video creation. Canas allows users to edit images, while Reel can create short videos with adjustable camera motions. Both models include features for responsible use. AWS plans to expand Nova's capabilities, including a speech-to-speech model in early 2025 and an "any-to-any" model later that year. These developments aim to enhance interactions across various media formats, including text, speech, images, and video.