Supported models and regions for Amazon Bedrock knowledge bases - Amazon Bedrock
Supported models for vector embeddingsSupported models and Regions for parsingSupported models and Regions for reranking results during query

Supported models and regions for Amazon Bedrock knowledge bases

Amazon Bedrock Knowledge Bases is supported in the following Regions (for more information about Regions supported in Amazon Bedrock see Amazon Bedrock endpoints and quotas):

  • US East (N. Virginia)

  • US East (Ohio)

  • US West (Oregon)

  • AWS GovCloud (US-West)

  • Asia Pacific (Tokyo)

  • Asia Pacific (Seoul)

  • Asia Pacific (Mumbai)

  • Asia Pacific (Singapore) (Gated)

  • Asia Pacific (Sydney)

  • Canada (Central)

  • Europe (Frankfurt)

  • Europe (Zurich)

  • Europe (Ireland) (Gated)

  • Europe (London)

  • Europe (Paris)

  • South America (São Paulo)

You can use the following foundation models (to see which Regions support each model, refer to Supported foundation models in Amazon Bedrock) for knowledge base query:

  • AI21 Labs Jamba 1.5 Large

  • AI21 Labs Jamba 1.5 Mini

  • AI21 Labs Jamba-Instruct

  • Amazon Titan Text G1 - Premier

  • Anthropic Claude 2.1

  • Anthropic Claude 2

  • Anthropic Claude 3 Haiku

  • Anthropic Claude 3 Sonnet

  • Anthropic Claude 3.5 Haiku

  • Anthropic Claude 3.5 Sonnet v2

  • Anthropic Claude 3.5 Sonnet

  • Cohere Command R+

  • Cohere Command R

  • Meta Llama 3 70B Instruct

  • Meta Llama 3 8B Instruct

  • Meta Llama 3.1 405B Instruct

  • Meta Llama 3.1 70B Instruct

  • Meta Llama 3.1 8B Instruct

  • Meta Llama 3.2 11B Instruct

  • Meta Llama 3.2 90B Instruct

  • Mistral AI Mistral Large (24.02)

  • Mistral AI Mistral Large (24.07)

  • Mistral AI Mistral Small (24.02)

Amazon Bedrock Knowledge Bases also supports the use of inference profiles for parsing data or when generating responses. With inference profiles, you can track costs and metrics, and also do cross-region inference to distribute model inference requests across a set of regions to allow higher throughput and facilitate greater resilience. You can specify an inference profile in a RetrieveAndGenerate or CreateDataSource request. For more information, see Set up a model invocation resource using inference profiles.

Important

If you use cross-region inference, your data can be shared across regions.

You can also use SageMaker AI models or custom models that you train on your own data.

Note

If you use an SageMaker AI or custom model, you must specify the orchestration and generation prompts (for more information, see Knowledge base prompt templates in Configure and customize queries and response generation). Your prompts must include information variables to access the user's input and context.

Region and model support differ for some features in Amazon Bedrock Knowledge Bases. Select a topic to view support for a feature:

Supported models for vector embeddings

Amazon Bedrock Knowledge Bases uses an embedding model to convert your data into vector embeddings and store the embeddings in a vector database. For more information, see Turning data into a knowledge base.

Embedding models support the following vector types.

Model name Supported vector type
Amazon Titan Embeddings G1 - Text Floating-point
Amazon Titan Text Embeddings V2 Floating-point, binary
Cohere Embed (English) Floating-point, binary
Cohere Embed (Multilingual) Floating-point, binary

Supported models and Regions for parsing

When converting data into vector embeddings, you have different options for parsing your data in Amazon Bedrock Knowledge Bases. For more information, see Parsing options for your data source.

The following lists support for parsing options:

  • The Amazon Bedrock Data Automation parser is supported in US West (Oregon) and is in preview and subject to change.

  • The following foundation models can be used as a parser:

    • Anthropic Claude 3.5 Sonnet

    • Anthropic Claude 3 Haiku

Supported models and Regions for reranking results during query

When retrieving knowledge base query results, you can use a reranking model to rerank results from knowledge base query. For more information, see Query a knowledge base and retrieve data and Query a knowledge base and generate responses based off the retrieved data.

For a list of models and Regions that support reranking, see Supported Regions and models for reranking in Amazon Bedrock.