What would you like to see?
OpenRouter’s Responses API (which does not work with the generic OpenAI provider) seems to allow configurable reasoning and effort levels as well as some other potentially useful options. Currently reasoning can’t be changed in any way through AnythingLLM, and Claude models don’t support reasoning at all there.
https://openrouter.ai/docs/api/reference/responses/overview
What would you like to see?
OpenRouter’s Responses API (which does not work with the generic OpenAI provider) seems to allow configurable reasoning and effort levels as well as some other potentially useful options. Currently reasoning can’t be changed in any way through AnythingLLM, and Claude models don’t support reasoning at all there.
https://openrouter.ai/docs/api/reference/responses/overview