At the AWS re:Invent 2024 conference in Las Vegas, Amazon CEO Andy Jassy took to the stage to launch its suite of AI foundation models called Nova which enable users to generate text, images and videos.
Jassy said Nova is a significant step in the company’s AI journey as it brings innovative generative AI tools to businesses and individuals with significant cost savings and reduced latency.
Amazon Nova’s “secret sauce” is that it enables users to analyse complex documents and videos, understand charts and diagrams, generate video content, and build sophisticated AI agents.
The Amazon Nova suite comes in four options:
Nova Micro: A text-only model that delivers low latency responses. It is designed for tasks such as text summarisation, translation, content classification, interactive chat and brainstorming, and simple mathematical reasoning and coding.
Nova Lite: A multimodal model that is lightning fast for processing image, video, and text inputs to generate text output. It can handle real-time customer interactions, document analysis, and visual question-answering tasks with high accuracy. It can analyse multiple images or up to 30 minutes of video in a single request.
Nova Pro: A multimodal model with the best combination of accuracy, speed, and cost for a wide range of tasks. It demonstrates strong capabilities in processing both visual and textual information and excels at analysing financial documents. It can process code bases with over 15,000 lines of code. Amazon Nova Pro also serves as a teacher model to distil custom variants of Amazon Nova Micro and Lite.
Nova Premier: (to be added to the portfolio in early 2025): Used for complex reasoning tasks and as the best teacher for distilling custom models.
Amazon also intends to roll out two other models in 2025 – Nova Canvas and Nova Reel.
With Nova Canvas users will be able to edit images through natural language text prompts and adjust layouts or colour schemes. Built-in safety measures, such as watermarking and content moderation, ensure responsible AI usage.
Nova Reel is a video generation model supporting advanced features, including camera motion controls such as panning, zooming, and 360-degree rotations. It allows for the creation of dynamic six-second videos, with additional functionalities expected in the future.
Nova Canvas will provide image generation model producing studio-quality images with precise control over style and content, including rich editing features such as inpainting, out painting, and background removal.
Nova Reel will enable users to produce short videos through text prompts and images, control visual style and pacing, and generate professional-quality video content for marketing, advertising, and entertainment.
All the Nova models include built-in safety controls and creative content generation models include watermarking capabilities to promote responsible AI use.
According to Jassy, the models deliver significant advances in latency and cost savings. The Nova models are integrated into Bedrock, its own machine learning platform that simplifies access to high-performance AI models through a single API.
This means Nova users will need an Amazon Bedrock account to access them which is available here.
Nova has been trialled by a number of Amazon clients including Dentsu Digital and SAP.
Dentsu is integrating Amazon Nova Reel into its creative process, enabling its teams to improve and accelerate the development of its campaigns from briefing, to concept development, to creative video content generation. It believes it will reduce the overall time it takes to generate assets from weeks to days.
Saturo Yamamoto, Executive Officer at Dentsu, said: “At Dentsu our quest for innovation in digital marketing is constantly fuelled by leveraging cutting-edge technology, and Amazon Nova Reel video generation AI is helping us do just that. It empowers our teams to explore creative avenues more freely, turning what used to take weeks into days. With Amazon Nova, we rapidly create mock-ups and precise proposal scenarios while crafting short, impactful videos - a transformative change that has boosted our efficiency.”
Jassy added during his presentation: "We prioritise technology that we think will really matter for customers and with the explosion of generative AI over the last couple of years we have taken the same approach. There is a tonne of innovation, what we are trying to do is solve problems for you, what we think of as practical AI.”
In closed talking about the future of Nova: "So, what is going to be next for us in Nova? The team is going to be working really hard next year on the second generation of these models, but I also have a couple of things that I am going to give you a sneak peek into," he said as he hinted to the availability of a Nova speech-to-speech model and a Nova Any-To-Any both in 2025.
