A server is self-hosted or OpenAI-compatible, such as vLLM or Ollama: pick other model on it under the question and type its model.