Claude replies it is 3.5 / 3.7 / 4, while I am always using claude-4-sonnet-thinking

Discussions
1

I am using 100% claude-4-sonnet-thinking in MAX mode.

I have the following cursor rule:

  • ALWAYS begin your response by telling me what AI model you are using. Just say the name of the model like (claude-x-sonnet or gemini-x-pro) - replace the x for your actual version.

I am getting the following response back:

I’m using Claude 3.5 Sonnet. (40% of the time)
I’m using Claude 3.7 Sonnet. (50% of the time)
I’m using Claude 4.0 Sonnet. (10% of the time)

While all my requests are being logged as claude-4-sonnet-thinking in MAX mode.

Is this an issue with Claude that it doesn’t know which model it is or…?

2

Here are two longer now hidden/de-listed threads where it has been discussed. Cursor staff says everything is good, but I’m still getting poor results and it still says it is 3.5 with a knowledge cutoff that matches 3.5.

1 Like
3

hmmmmm… maybe time to switch to Claude Code or another competitor if they can’t make this transparent and working like it should.

1 Like