Retail giant has introduced Nova—a new family of foundational AI models capable of generating text, images, and videos at AWS re:Invent, the company’s annual conference.
According to the company’s blog post, the Nova suite includes four text-generating models: Micro, Lite, Pro, and Premier. Nova Micro is a text-to-text model, while Nova Lite, Pro, and Premier are multi-modal models capable of processing text, images, and videos to generate text.
Multi-model here refers to the ability of the model to process multiple types of inputs, spanning from text, images, audio, or video, and generate an output based on that combination.
Additionally, the tech behemoth has introduced two specialised models, Nova Canvas and Nova Reel, designed to create studio-quality images and videos, respectively.
“Inside Amazon, we have about 1,000 generative AI applications in motion, and we’ve had a bird’s-eye view of what application builders are still grappling with,” said Rohit Prasad, SVP of Amazon Artificial General Intelligence.
“Our new Amazon Nova models are intended to help with these challenges for internal and external builders, and provide compelling intelligence and content generation while also delivering meaningful progress on latency, cost-effectiveness, customisation, Retrieval Augmented Generation (RAG), and agentic capabilities,” he added.
Nova Micro, Lite, and Pro are up to 75% more cost-efficient compared to the top-performing models in their respective categories within Amazon Bedrock, said the blog.
The newly launched AI offerings have already gained traction from several large enterprises such as SAP, Deloitte, and Dentsu Digital, among others.
For instance, Musixmatch, a lyrics platform with over 80 million users, has integrated Amazon Nova Reel into its Musixmatch Pro offering, to allow creators to distribute lyrics across major digital streaming platforms and social networks. Budding artists can utilise the ‘Nova Reel’ feature to create high-quality music videos by using their song’s context as input, along with customising them with prompts.
