[AWS Bedrock] Claude 3.5 API - want to apply max output token to header

Question

I'm using AWS Bedrock.

I checked the article about the increase in the number of output tokens.
- link: https://docs.anthropic.com/en/docs/about-claude/models

**"8192 output tokens is in beta and requires the header anthropic-beta: max-tokens-3-5-sonnet-2024-07-15. If the header is not specified, the limit is 4096 tokens."**

How can I apply the above code in a jupyter file using AWS Bedrock?

I applied this like below. 
![Enter image description here](/media/postImages/original/IMoOu9oT32QAC0pYoelEMKIg)
![Enter image description here](/media/postImages/original/IMKADwoz2BRvC8T20XiVkkvQ)

I saw the following error phrase.
![Enter image description here](/media/postImages/original/IMKiGxm1QeTICn5lOy19IAbg)

Didier_Durand · Answer

Hi,

Sonnet v3.5 2024-07-15 is probably not yet activated on Bedrock.

If you look at this page: https://docs.aws.amazon.com/bedrock/latest/userguide/model-ids.html , the id of Sonnet v3.5 is `anthropic.claude-3-5-sonnet-20240620-v1:0`

So, June version instead of the July version that you are looking for. You'll have to wait for an upgrade, I guess.

Best,

Didier

rePost-User-0204871 · Answer

Check out V2.  I tested Cloud Sonnet 3.5 V2 today and it works fine with the 8K output tokens.

[AWS Bedrock] Claude 3.5 API - want to apply max output token to header

Relevant content