OpenAI presents developers with advanced voice and reasoning models

OpenAI has announced several new options for developers who use its technology to build products and services, promising the improvements will “improve performance, flexibility, and cost-effectiveness.”

In their live announcement today — with audio issues — the OpenAI team first highlighted changes to OpenAI o1, the company’s reasoning model that “can handle complex multi-step tasks,” according to the company. Developers can now use the model in their advanced application class; it is currently used by engineers to build automated customer service systems, to help inform supply chain decisions, and to predict financial trends.

The new o1 model can also connect to external data and APIs (also called Application Programming Interfaces, which is how different software applications communicate with each other). Developers can also use o1 to fine-tune messages to give their AI applications a specific tone and style; the model also has the ability to see to use images to “open up many applications in science, manufacturing, or coding, where visual input is important.”

Improvements were also announced to OpenAI’s Realtime API, which developers use for voice assistants, virtual tutors, translation bots, and AI Santa voices. The company’s new WebRTC support will help with real-time voice services, using JavaScript to make transparently better sound quality and useful responses (eg, the RealTime API can start generating responses to a question even while the user is still speaking). OpenAI also announced price reductions for services such as WebRTC Support.

And importantly, OpenAI now offers Preference Fine-Tuning to developers, which customizes the technology to respond better to “primary tasks where tone, style, and creativity are important” than so-called Supervised Direct Tuning. Catch the full presentation below.