Nova Pro 1.0

Nova Pro 1.0 - Multimodal understanding model

Nova Pro 1.0 is Amazon’s high-capability multimodal model for balanced accuracy, speed, and cost, with support for long-document and visual workflows.

Defining Characteristics

• 300K-token context window: AWS and provider docs list a 300,000-token context window for long prompts, codebases, and multi-turn analysis.
• Multimodal inputs: The model accepts text and image/vision inputs; AWS positions the Nova family across text, image, video, speech, API calling, and agentic AI capabilities.
• Tool use and function calling: Official and provider docs explicitly mention tool use and function calling for agent-augmented workflows.
• Fine-tuning support: Nova Pro supports fine-tuning on Amazon Bedrock, allowing adaptation for specific domains and cost targets.

Core Design

• Amazon describes Nova Pro as the highest-capability understanding model in the Nova family for tasks that need accuracy, response speed, and cost balance.
• Official docs also frame it as suitable for agentic workflows and UI actuation.