OpenAI introduces API for Whisper speech-to-text technology and ChatGPT

Developers can now officially incorporate ChatGPT into their products and services.

OpenAI has introduced an API that allows third-party developers to integrate ChatGPT into their apps and services at a significantly lower cost than using existing language models. Whisper, an AI-powered speech-to-text model, is also available for use through an API with some key changes to OpenAI's developer terms of service.

While ChatGPT is commonly used for creating AI-powered chat interfaces, OpenAI states that its new model family, gpt-3.5-turbo, is the "best model for many non-chat use cases." However, this model is not expected to be the same as the one Microsoft's Bing is using.

OpenAI offers 1,000 tokens for $0.002, which is 10x cheaper than its existing GPT-3.5 models. It provides a tool for determining the number of tokens needed to interpret text and notes that one token corresponds to around four characters in English.

OpenAI has also launched another API for Whisper, its speech-to-text model, which can transcribe or translate audio at a cost of $0.006 per minute. Whisper is open source, but OpenAI's powerful hardware allows for a faster turnaround, making the API a suitable option for those seeking quick transcription on low-powered devices.

OpenAI has implemented policy changes in response to developer feedback, such as moving from an opt-out to an opt-in system for using data submitted through the API to train its models. Additionally, the company is working on improving uptime and prioritizing the stability of production use cases.

Overall, OpenAI's new ChatGPT API is expected to provide an affordable and accessible solution for developers seeking to integrate AI chatbots into their apps and services.