Openai推出gpt-live即時語音api任何人都能構建低延遲智能體
openai發布了gpt-live，這是一個低延遲、無輪流等待的語音模型和系統架構，專為連續語音互動而設計。這使得開發者能夠構建更自然、響應更快的語音智能體，消除輪流對話帶來的摩擦。
openai released technical details on gpt-live, a realtime voice system developed in six months that utilizes a turnless speech model to facilitate continuous conversation. the architecture prioritizes low latency to create faster, more natural interactions by processing audio streams without waiting for user pauses. this upgrade targets developers building next-gen voice agents, offering a path to more fluid experiences compared to traditional request-response models. it indicates a broader shift in the industry towards multimodal, realtime interfaces as the primary interaction method for advanced ai systems.