Nova Models by Amazon Web Services (AWS) AI
Details
Capabilities
Links
WebsiteAbout
AWS Nova is a new family of foundation models introduced by Amazon Web Services, focusing on delivering cutting-edge generative AI capabilities across text, multimodal content, and video creation. Designed to support a wide range of enterprise applications, Nova models emphasize scalability, efficiency, and customization. With industry-leading price-performance, Nova enables organizations to integrate advanced generative AI into their workflows, unlocking opportunities in automation, creativity, and data insights. The Nova suite leverages AWS's robust infrastructure, including fine-tuning options and multimodal intelligence. This includes features that enable high-quality text generation and the creation of sophisticated visual content. Nova integrates seamlessly with Amazon Bedrock, allowing businesses to implement safeguards, evaluate models using custom metrics, and optimize generative AI applications for responsible and reliable outcomes. AWS Nova is part of AWS's broader push to democratize generative AI by making powerful and versatile models accessible to diverse industries. By incorporating advanced evaluation tools and built-in safety mechanisms, AWS ensures that Nova models can address both technical and ethical challenges, enabling enterprises to innovate responsibly at scale.
Current Variants
Use-when guidance is based on each model's tracked capabilities, context window, release date, and replacement status.
Use when the workload needs multimodal inputs.
Use when the workload needs realtime voice and audio.
Use when provider availability and model metadata match the workload.
Use when the workload needs 128k context and structured outputs.
Use when the workload needs 300k context, structured outputs, and multimodal inputs.
Use when the workload needs 300k context, structured outputs, and multimodal inputs.
Use when the workload needs 1m context, reasoning, and tool use.
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| Amazon Nova Multimodal Embeddings | Use when the workload needs multimodal inputs. | 2025-12 | multimodal inputs | Current |
| Amazon Nova 2 Sonic | Use when the workload needs realtime voice and audio. | 2025-11 | realtime voiceaudio | Current |
| Amazon Nova Canvas | Use when provider availability and model metadata match the workload. | 2025-11 | — | Current |
| Amazon Nova Reel | Use when provider availability and model metadata match the workload. | 2025-11 | — | Current |
| Nova Micro | Use when the workload needs 128k context and structured outputs. | 2025-03 | 128k contextstructured outputs | Current |
| Nova Lite | Use when the workload needs 300k context, structured outputs, and multimodal inputs. | 2025-03 | 300k contextstructured outputsmultimodal inputs | Current |
| Nova Pro | Use when the workload needs 300k context, structured outputs, and multimodal inputs. | 2025-03 | 300k contextstructured outputsmultimodal inputs | Current |
| Amazon Nova Premier | Use when the workload needs 1m context, reasoning, and tool use. | 2025-03 | 1m contextreasoningtool use | Current |
Release Timeline
3 release groupsSpecifications(8 models)
| Model | Released | Context | Vision | Multimodal | Reasoning | Fn Calling | Tool Use | Structured Outputs |
|---|---|---|---|---|---|---|---|---|
| Amazon Nova Multimodal Embeddings | 2025-12 | — | No | Yes | No | No | No | No |
| Amazon Nova 2 Sonic | 2025-11 | — | No | No | No | No | No | No |
| Amazon Nova Canvas | 2025-11 | — | No | No | No | No | No | No |
| Amazon Nova Reel | 2025-11 | — | No | No | No | No | No | No |
| Nova Micro | 2025-03 | 128k | No | No | No | No | No | Yes |
| Nova Lite | 2025-03 | 300k | Yes | Yes | No | No | No | Yes |
| Nova Pro | 2025-03 | 300k | Yes | Yes | No | No | No | Yes |
| Amazon Nova Premier | 2025-03 | 1m | Yes | Yes | Yes | Yes | Yes | No |
Available From(2 providers)
Pricing
| Model | Provider | Input / 1M | Output / 1M | Type |
|---|---|---|---|---|
| Nova Micro | OpenRouter | $0.035 | $0.14 | Serverless |
| Nova Micro | AWS Bedrock | $0.04 | $0.14 | Serverless |
| Nova Lite | OpenRouter | $0.06 | $0.24 | Serverless |
| Nova Lite | AWS Bedrock | $0.06 | $0.24 | Serverless |
| Nova Pro | OpenRouter | $0.8 | $3.2 | Serverless |
| Nova Pro | AWS Bedrock | $0.8 | $3.2 | Serverless |
| Amazon Nova Premier | OpenRouter | $2.5 | $12.5 | Serverless |
| Amazon Nova Premier | AWS Bedrock | $2.5 | $12.5 | Serverless |
Popular comparisons in this family
Frequently Asked Questions
- What is Nova used for?
- Nova is used for realtime voice, vision and multimodal work, and reasoning. The family description and listed model capabilities point to those workloads as the best fit.
- How does Nova compare to Amazon Nova?
- Nova by Amazon Web Services (AWS) AI is strongest where you need realtime voice, while Amazon Nova by Amazon Web Services (AWS) AI is the closest related family to check for audio. Nova has 8 listed variants and reaches up to 1m context, while Amazon Nova reaches up to 1m context, so compare the specs and pricing tables before choosing a production model.
- Which Nova model should I use?
- For the lowest listed input price, start with Nova Micro through OpenRouter at $0.035/1M input tokens. For the most capable/latest local choice, evaluate Amazon Nova Premier with 1m context and reasoning, tool use, function calling, and multimodal inputs.
Related Model Families
Models(8)
Amazon Nova Multimodal Embeddings
Amazon Nova 2 Sonic
Amazon Nova Canvas
Amazon Nova Reel
Nova Micro
Nova Lite
Nova Pro
Amazon Nova Premier