Run AIMLAPI LLM and reasoning workflows through chat completions with retries, structured outputs, and explicit User-Agent headers. Use when Codex needs scripted prompting/reasoning calls against AIMLAPI models.

通过聊天补全功能运行AIMLAPI的LLM和推理工作流，支持重试、结构化输出和显式的User-Agent头。当Codex需要对AIMLAPI模型进行脚本化提示或推理调用时使用。