Generative AI
Overview Our Story Customers Partners

HAQM Nova Understanding Models

Frontier intelligence and industry leading price performance

Get started with HAQM Nova
Learn more at the user guide

HAQM Nova understanding models

HAQM Nova Micro, HAQM Nova Lite, and HAQM Nova Pro are state-of-the-art models that deliver frontier intelligence and industry-leading price performance. 

  • HAQM Nova Micro is a text-only model that delivers the lowest latency responses at a very low cost. 
  • HAQM Nova Lite is a very low-cost multimodal model that is lightning fast for processing image, video, and text inputs to generate text outputs.
  • HAQM Nova Pro is a highly capable multimodal model with the best combination of accuracy, speed, and cost for a wide range of tasks. 

HAQM Nova understanding models support over 200 languages, text and vision fine-tuning, and easy integration with proprietary data and applications through HAQM Bedrock features such as HAQM Bedrock Knowledge Bases and HAQM Bedrock Agents. Using the appropriate model and techniques like distillation, you can get the optimal intelligence, latency, and costs for your needs with HAQM Nova understanding models. 

Key capabilities

HAQM Nova understanding models deliver state-of-the-art text and visual intelligence, with native support for plain text, documents, image, and video understanding. 

HAQM Nova Canvas and HAQM Nova Reel allow multiple ways for you to control and customize the generated content. You can provide a reference image to guide your image or video generation. HAQM Nova Canvas also lets you use text inputs to determine the color scheme or “color palette” for the image. With camera motion control in HAQM Nova Reel, you can guide camera motion with text prompts like “zoom” and “dolly forward” to get the exact visual needed for your video.

HAQM Nova Lite, Micro, and Pro are among the fastest and most cost-effective models in their respective intelligence classes. With new latency-optimized inference options, enterprises can achieve even faster response times for their mission-critical AI applications without compromising quality.

HAQM Nova understanding models excel at in-context learning (ICL) and grounding responses based on contextual data provided in approaches like Retrieval Augmented Generation (RAG). This is seen in HAQM Nova model scores in the Comprehensive RAG Benchmark (CRAG) evaluation. You can easily integrate HAQM Nova understanding models with their proprietary data and applications using HAQM Bedrock features such as HAQM Bedrock Knowledge Bases.

New generative AI applications are increasingly using AI agents to interact with external services and APIs and to perform tasks in the customer’s environment. HAQM Nova models excel at agentic workflows and are the state of the art on benchmarks that evaluate this capability, like Berkeley Function Calling Leaderboard (BFCL), VisualWebBench, and Mind2Web.
 
  • HAQM Nova Micro
  • HAQM Nova Lite
  • HAQM Nova Pro

The numbers for other models are a mix of self-reported and measured evaluations. For a more detailed overview of evaluations, refer to the HAQM Nova technical report.

HAQM Nova Pro in action

In this demo, we show how HAQM Nova models can power computer Agents, which plan and execute multi-step actions, using the browser and screen-based user interfaces as universal tools to automate tasks for end customers. Let's see an example of how a sales representative can interact with a computer agent powered by HAQM Nova to use SAP S/4Hana. SAP is one of the most widely used SaaS ERP systems.