<p>OpenAI has introduced gpt-realtime-1.5, an update to their Realtime API, which significantly enhances voice workflow capabilities. This advanced model improves the reliability of instruction following, enhances tool calling, and offers superior multilingual accuracy.</p>
<p>Designed specifically for voice agents, the gpt-realtime-1.5 facilitates low-latency, speech-to-speech interactions, allowing seamless integration with various multimodal inputs such as audio, images, and text. Outputs can be generated in both audio and text formats, making it a versatile tool for a diverse range of applications. These advancements position gpt-realtime-1.5 as a powerful upgrade for developers seeking to leverage cutting-edge AI in their voice processing and interaction workflows.</p>


Collections

OpenAI has released gpt-realtime-1.5, an update to its Realtime API targeting voice agent development. The new model improves instruction-following reliability, tool calling, and multilingual accuracy. It supports low-latency speech-to-speech interactions and integrates with multimodal inputs including audio, images, and text, with outputs available in both audio and text formats.

Enhancements in Voice Workflows with gpt-realtime-1.5 API