Amazon.com Inc has announced new regional processing options for its Amazon Nova understanding models in the European Union and Asia Pacific. These models, part of a new generation of foundation models, offer state-of-the-art intelligence across various tasks with competitive price performance.
AWS customers can now use Cross-Region inference profiles for Amazon Nova understanding models—Amazon Nova Lite, Amazon Nova Micro, and Amazon Nova Pro—in several regions. These include the European Union (Stockholm, Frankfurt, Ireland, Paris) and Asia Pacific (Tokyo, Seoul, Mumbai, Singapore, Sydney).
Cross-region inference profiles allow automatic routing of requests to Amazon Nova across multiple regions while prioritizing the source region to reduce latency. This method is cost-efficient as there are no extra routing costs; charges depend solely on the request's source region.
The available models include:
- Amazon Nova Micro: A text-only model offering low latency responses at minimal cost.
- Amazon Nova Lite: A low-cost multimodal model for processing image, video, and text inputs swiftly.
- Amazon Nova Pro: A multimodal model combining accuracy, speed, and cost efficiency for various tasks.
These models support over 200 languages and enable integration with proprietary data through features like Amazon Bedrock Knowledge Bases.
“Inside Amazon, we have about 1,000 generative AI applications in motion," stated Rohit Prasad, SVP of Amazon Artificial General Intelligence. "Builders across every Amazon business are using our new Amazon Nova models to help with these challenges today."
All Amazon Nova models are designed to be fast and cost-effective. They are at least 75 percent less expensive than top-performing models in their intelligence classes within the AWS US regions. Integration with Amazon Bedrock allows customers to experiment with these models through a single API.
Amazon emphasizes that its Nova models excel at Retrieval Augmented Generation (RAG), ensuring response accuracy by grounding them in an organization's data. The models are optimized for agentic applications requiring interaction with proprietary systems through multiple APIs.
Several AWS partners and customers are already utilizing these capabilities:
- SAP is integrating these models into its SAP AI Core hub to enhance automation and personalization.
- Deloitte highlights advanced customization capabilities driving innovation for global clients.
- Palantir Technologies integrates advanced reasoning capabilities within its Ontology System.
- Cognigy transforms customer service experiences using Agentic AI powered by Amazon Nova Pro.
- Ativion’s StudentKeeper uses AI-assistant features for digital classroom management.
Amazon remains committed to responsible AI development by embedding safety measures in its Nova models and launching AWS AI Service Cards detailing use cases and limitations.
For more information on starting with Amazon Nova models: https://aws.amazon.com/nova/