Introducing the Anthropic Claude 3 Sonnet Model on Amazon Bedrock
The Claude 3 Sonnet model by Anthropic is a cutting-edge large language model designed to balance speed and intelligence. This model supports both text and image input, producing text outputs, and boasts an impressive 200,000 token context window, making it ideal for handling lengthy documents and complex tasks.
Availability
The Claude 3 Sonnet model is currently available in the US East (N. Virginia) and US West (Oregon) regions on Amazon Bedrock, with plans to expand to additional regions soon. Note that some users have reported availability issues despite the model being listed in their AWS accounts.
Usage
To access the model, you can use the Amazon Bedrock console, AWS CLI, or AWS SDKs like Python/Boto3. The model ID for this version is anthropic.claude-3-sonnet-20240229-v1:0.
Features and Capabilities
- Context Window: 200,000 tokens
- Input/Output: Supports text and image input, with text output
- Performance: Balanced intelligence and speed, suitable for various applications
Pricing
The Claude 3 Sonnet model is available on a pay-as-you-go basis with options for On-Demand and Batch mode, as well as Provisioned Throughput mode. For detailed pricing information, refer to the Amazon Bedrock pricing documentation.
Technical Details
Integration with the model is possible through various APIs and SDKs. For instance, the Spring AI documentation provides guidance on using the model programmatically with Java. When making API calls, you can specify the model ID and other configuration options like temperature, top-p, top-k, and max tokens.
Support and Troubleshooting
If you encounter issues with model availability or usage, it is recommended to contact AWS Support for assistance. While there have been some initial rollout bugs and errors, these are actively being addressed.
