而这次的GPT-Realtime-1.5，更是把这种体验推向了极致。它变得更加懂人话了，甚至能听出你语气里的急躁或者犹豫，那种机器味淡了很多，感觉就像是在跟一个真人打电话。 说到实际应用，这次更新最大的受益者就是语音智能体。你可以用它来开发智能客服、语音助手，甚至是电话销售机器人。

在模型方面，全新的实时模型gpt-realtime-1.5及其配套音频模型已正式发布。它们的核心目标是提高语音指令的可靠性。根据OpenAI的内部测试数据，新模型在数字和字母的转录准确率方面提高了约10%，逻辑音频任务的准确率提高了5%，指令执行的准确率也提高了7%，有效解决了AI在听取关键短语或执行复杂语音指令时出现偏差的问题。

OpenAI 指出，这一改进对于需要频繁调用大量工具的复杂 AI 代理尤为关键，能够将其运行速度直接提升 20% 到40% 。这两项更新不仅让 AI 的“听力”更敏锐，更让其“行动”效率迈向了全新的台阶。

The OpenAI ChatGPT Realtime API, now available in public beta, is transforming how developers create low-latency, multimodal applications. By seamlessly integrating speech, text, and function calling ...

Nearly a year after the developer preview was introduced, OpenAI released the GA version (General Availability) of the Realtime API in August 2025. The Realtime API is a multimodal interface that ...

OpenAI has introduced the public beta of its Realtime API, offering developers a tool to integrate natural, low-latency, multimodal interactions into their applications. Now available to all paid ...

OpenAI’s new GPT-Realtime model and Realtime API updates bring lifelike voice AI, phone calling, and image input to everyday apps. If you’ve ever wished that talking to an AI felt more like chatting ...

Realtime API supports multi-model text and speech experiences including natural speech-to-speech conversations using preset voices already supported in the API. OpenAI has introduced a public beta of ...

OpenAI has unveiled its latest speech-to-speech artificial intelligence (AI) model, gpt-realtime, designed to generate more vivid and natural voice interactions for real-time applications. Alongside ...

As more companies integrate large language models into customer support, analytics, and internal automation, the main concern ...

Want smarter insights in your inbox? Sign up for our weekly newsletters to get only what matters to enterprise AI, data, and security leaders. Subscribe Now As enterprise developers and astute company ...

OpenAI said on Monday that it would soon wind down the availability of GPT-4.5, its largest-ever AI model, via its API. GPT-4.5 was released only in late February. Developers will have access to GPT-4 ...