一句话摘要
GPT Realtime 和 Audio 模型 GA,新增图像输入、异步 Function Calling、对话模式等,语音 Agent 能力全面成熟。
详细描述
OpenAI's GPT RealTime and Audio models are now generally available in Microsoft Foundry. Improvements include enhanced instruction following, new standard voices (Marin, Cedar), improved audio quality, image input support, improved async function calling, and Conversation Mode with VAD.
原文摘录
OpenAI's GPT RealTime and Audio models are now generally available in Microsoft Foundry Models. Improvements: Enhanced instruction following, new standard voices Marin and Cedar, improved audio quality, Image Input support, improved function calling with async support, Conversation Mode with VAD.