Amazon has launched new generative AI-powered foundational models known as ‘Amazon Nova’ on Amazon Bedrock.
The “new generation of foundation models” are capable of processing text, images, and videos as prompts, understanding videos, charts, and documents, and generating videos and other content.
According to the biggest online retailer and technology provider, Amazon Nova can provide customers with frontier intelligence and “industry-leading” price performance.
Rohit Prasad, SVP of Amazon Artificial General Intelligence, outlines some of the potential use cases for Amazon Nova:
“Inside Amazon, we have about 1,000 Gen AI applications in motion, and we’ve had a bird’s-eye view of what application builders are still grappling with.
“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, information grounding, and agentic capabilities.”
Amazon Nova Models
The new Amazon Nova models include a text-only model Amazon Nova Micro, a low-cost multimodal model Amazon Nova Lite, an advanced multimodal model Amazon Noa Pro, the best multimodal model Amazon Nova Premier, an image generation model Amazon Nova Canvas, and a video generation model Amazon Noa Reel.
Amazon Nova Premier will be made available in the first quarter of 2025.
These models support a long list of tasks across 200 languages and various modalities.
Amazon asserts than its Amazon Nova Micro, Amazon Nova Lite, and Amazon Nova Pro are a minimum of 75 percent cheaper than the best performing models within their intelligence classes on Amazon Bedrock.
The fully managed service Amazon Bedrock offers a range of its own high-performance foundation models along with those from leading artificial intelligence companies, such as Anthropic, Cohere, Meta, AI21 Labs, Mistral AI, Stability AI, and more.
Customers can use Amazon Bedrock to test and evaluate the different Amazon Nova models to work out which is the best fit for them.
The models can also be customised to use the customer’s proprietary data ni order to improve accuracy and personalisation.
In Amazon’s related press release, it offers an insights into its future plans:
“We will introduce two additional Amazon Nova models in 2025, including a speech-to-speech model and a native multimodal-to-multimodal—or “any-to-any” modality model.
“Our speech-to-speech model will understand streaming speech input in natural language, interpreting verbal and nonverbal cues (like tone and cadence), and delivering natural humanlike interactions, while our any-to-any model will be capable of processing text, images, audio, and video, as both input and output.
“It will simplify the development of applications where the same model can be used to perform a wide variety of tasks, such as translating content from one modality to another, editing content, and powering AI agents that can understand and generate all modalities.”
Amazon reassures customers that its Amazon Nova models have been built with safety measures and protections. It has also launched AWS AII Service Cards for these models, providing information on use cases, limitations, and its approach to responsible AI.
Amazon has been on centre stage for AI news over the past week, It has unveiled new datacentre components that will ultimately be able to power more customer AI solutions.
AWS IoT SiteWise announced the general availability of a new generative AI-powered industrial assistant, which offers insights and problem-solving capabilities following natural language prompts by the user.
