Description
Claude 3.7 Sonnet is a hybrid model, and has a thinking mode, that can be enabled using an API parameter (https://docs.anthropic.com/en/api/messages#body-thinking), and looks like RubyLLM doesn't support that just yet. The API also allows to be able to set a max output tokens and thinking budgets (where thinking budget is lower than max output tokens).
I can't find a way to enable these, making sure I'm not missing anything.
I'm also happy to spend time implementing this feature. I wanted to ask before I started any work on this to avoid duplicate work. I was thinking about starting with a slightly more open .with_arbitrary_options, that would take the options, and add it to the messages API.
Or if anyone else has ideas, happy to discuss/collab to get this going.