I believe open router maintains it. However for the gemini models, you cant set the low high, but you can adjust parameters such as sampling tempature, token limit, ect, to configure the models creativity, logic, and response.
It’s part of the official open source code base of n8n and since it is open source, anyone can maintain it with approval from the n8n team via a pull request.
Your options are:
to fork the n8n code base and add the missing options in the file below, or;
you could follow the documentation on contributing to the code base of n8n yourself for everyone to benefit.
If you’re not comfortable with coding yourself, then I guess you could open an issue on github for this