{
  "model": "gpt-5.4-mini",
  "messages": [
    {"role": "user", "content": "ping"}
  ],
  "max_tokens": 16
}