using claude sonnet 3.7 routing to claude 3 opus

It was good to see Claude 3.7 Sonnet recently released to AWS gov-cloud, as version 3.5 is not useful for software development.

However, I began to get suspicious when I started using it as it made the same mistakes as v3.5, hallucinating and failing miserably at anything and everything. So I asked it what version it is and it responded with:

Yes, this is Claude. I'm Claude, an AI assistant created by Anthropic. I'm currently running as Claude 3 Opus, Anthropic's most capable model.

This is with the "Chat / Text Playground" in the AWS console. I get the same thing when using VSCode+Continue. Did AWS deploy the wrong model to gov-cloud? I don't really care what it calls itself, but the vibe coding coming out of the model is identical to v3.5 when v3.7 is supposed to be "actually usable". Also:

My training data has a cutoff date in 2023

Has the wrong model been deployed?

There are a few inaccuracies in your statement that I'd like to address:

1. There is no model called "Claude Sonnet 3.7" or "Claude Opus 3" mentioned in the available information.

2. The most recent Claude models available on Amazon Bedrock are part of the Claude 3 family, which includes Claude 3 Opus, Claude 3 Sonnet, and Claude 3 Haiku.

3. Amazon Bedrock does offer an Intelligent Prompt Routing feature, which can route requests between different models based on the complexity of the prompt. However, this routing is designed to optimize for quality and cost, not to degrade performance.

4. If you're using a specific model, you should be billed for that model. The routing feature doesn't change your billing without your knowledge.

5. The model should not misidentify itself. If you're using Claude 3 Sonnet, it should identify as such, not as Opus.

If you're experiencing unexpected behavior or seeing inconsistent model identification, I would recommend:

1. Double-check which model you're actually invoking in your API calls or through the Amazon Bedrock console.
2. Review your Amazon Bedrock settings to see if you have Intelligent Prompt Routing enabled.
3. Check your billing details to confirm which model you're being charged for.
4. If the issue persists, contact AWS support for clarification and assistance.

It's important to use the correct model names and understand the features you're using to avoid confusion and ensure you're getting the expected service.
Sources
Reduce costs and latency with Amazon Bedrock Intelligent Prompt Routing and prompt caching (preview) | AWS News Blog
Unlocking Innovation: AWS and Anthropic push the boundaries of generative AI together | AWS Machine Learning Blog
Community | Mastering Amazon Bedrock with Claude 3: Developer's Guide with Demos