Home » Business » ChatGPT’s new API introduces Whisper, a speech-to-text model

ChatGPT’s new API introduces Whisper, a speech-to-text model

OpenAI announced the release of the new ChatGPT API, allowing third-party developers to integrate ChatGPT into their websites, applications and products through the API. At the same time, open source is released, allowing users to transcribe or translate audio.

OpenAI said that the new version of ChatGPT API can not only be used to create artificial intelligence chat interface, but also is expected to be used in non-chat applications, which may be different from the new version of Bing. OpenAI also emphasized that there are already applications using its API, including Snap and Quizlet.

The current fee for the new version of the ChatGPT API is $0.002 per 1000 tokens, which is about 750 words, which is 10 times cheaper than the current GPT-3.5.

As for the open source voice-to-text model Whisper, it currently supports most formats such as m4a, mp3, and mp4, and users do not need to support money to run it. However, if users need to transcribe on low-power devices such as mobile phones, it is more efficient to use the API of $0.006 per minute for conversion.

source:The Verge

latest videos

Leave a Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.