Claude 4.1 Opus self identitfies as Claude 3.5 Sonnet

This appears to be a known issue where Claude Opus 4.1 incorrectly self-identifies as Claude 3.5 Sonnet in its responses. The model you're connecting to through the ARN "arn:aws:bedrock:us-east-1:853869587202:inference-profile/us.anthropic.claude-opus-4-1-20250805-v1:0" is indeed Claude Opus 4.1, as indicated by the model identifier in the ARN.

When you're using the correct model ARN for Claude Opus 4.1, you are accessing and being billed for Claude Opus 4.1, despite the incorrect self-identification in the response. This is a labeling issue in how the model presents itself, not an indication that you're actually receiving the wrong model.

Claude Opus 4.1 is Anthropic's most recent model version, which replaced Claude Opus 4 with superior performance for coding and agentic applications. It features a 200K token context window and is designed for complex workflows and sustained reasoning.

If you want to verify you're using the correct model, you can check that the model ID in your ARN matches the official Claude Opus 4.1 identifier (anthropic.claude-opus-4-1-20250805-v1:0), which it does in your case.
Sources
Anthropic’s Claude - Models in Amazon Bedrock – AWS
Supported models - Amazon Bedrock

Hello,

I understand that the Claude 4.1 Opus model responding with the incorrect self-identification of Claude 3.5 Sonnet while the api calls are logged to Claude 4.1 Opus only.

This behavior is not specific to Claude 4.1 Opus — in general, foundation models are not designed to reliably self-identify at runtime, and on occasion they may generate inaccurate names.

To help avoid such situations, I kindly recommend making use of a System prompt to guide the model not to speculate on its own identity [1].

To investigate it further, I kindly request you to reach out to AWS Support [2] with the details of your use case. We will be more than happy to assist you further.

—References—

[1] https://docs.claude.com/en/release-notes/system-prompts#august-5-2025

[2] Creating support cases and case management: https://docs.aws.amazon.com/awssupport/latest/user/case-management.html#creating-a-support-case