Amazon Enhances AI Capabilities with Advanced 'Nova' Models
Click to start listening
Las Vegas, Dec 4 (NationPress) Amazon has introduced a groundbreaking series of artificial intelligence (AI) foundation models known as 'Nova', delivering top-tier intelligence across various tasks while maintaining industry-leading price performance.
These AI models will be integrated into the Amazon Bedrock model library on AWS, as announced during the annual 'AWS re:Invent' event.
“Within Amazon, we have around 1,000 generative AI applications in development, providing us with a comprehensive understanding of the challenges faced by application developers,” stated Rohit Prasad, SVP of Amazon Artificial General Intelligence.
The newly launched Amazon Nova models aim to address these challenges for both internal and external developers, offering advanced intelligence and content generation while significantly enhancing latency, cost-effectiveness, customization, Retrieval Augmented Generation (RAG), and agentic functions, Prasad elaborated.
As per the company, Amazon Nova Micro outperformed or matched both Meta LLaMa 3.1 8B across all 11 relevant benchmarks, and Google Gemini 1.5 Flash-8B on all 12 relevant benchmarks.
With an impressive speed of 210 output tokens per second, Amazon Nova Micro is perfectly suited for applications that demand rapid responses, the company stated.
In his keynote speech, AWS CEO Matt Garman emphasized the company's commitment to delivering innovative solutions in generative AI, including the new Trainium2 instances, Trainium3 chips, and the Amazon Nova foundation models.
The company announced plans to launch an Amazon Nova speech-to-speech model in the first quarter of 2025.
This model aims to revolutionize conversational AI applications by comprehending streaming speech input in natural language, interpreting both verbal and non-verbal signals (such as tone and cadence), and facilitating natural human-like interactions with minimal latency.
Amazon is also working on a unique model that can process text, images, audio, and video inputs to generate outputs in any of these formats.
This Amazon Nova model, featuring native multimodal-to-multimodal – or “any-to-any” modality capabilities – is expected to launch in mid-2025.
AWS also unveiled the next generation of Amazon SageMaker, consolidating essential features for rapid SQL analytics, petabyte-scale big data processing, data exploration and integration, model development and training, and generative AI into a single cohesive platform.
