Pretty much that. It would be useful to have an option at the API dashboard (and an API endpoint) to be able to set per-model or global (per-model having priority over global) settings like sampling options (temperature, top_p, etc), reasoning settings and venice parameters (especially include_venice_system_prompt).