Ever since the release of ChatGPT, developers have been trying to replicate its capabilities. Now, there is a good news for businesses and enterprises. OpenAI has released their chatbot’s API that users can now integrate into their apps and products. Not only ChatGPT’s cutting edge technology, developers can now also access the speech-to-text capabilities of Whisper.

By implementing a series of system-wide enhancements, OpenAI has reduced the cost of ChatGPT by 90% since December and are now passing on those savings to API users. The API runs on Microsoft Azure.

Snapchat, Quizlet, Instacart, Shopify, and Speak are some of the early adopters of the technology and have already implemented ChatGPT’s API on their platforms for providing their customers with recommendations, personalised assistance, and open-ended conversational abilities.

To get started, register here.

Sam Altman confirmed on Wednesday that to improve the model, the company won’t be using user submitted data anymore. “Any data submitted prior to March 1 is used for improvement,” and has a 30-day retention policy to identify any abuse or misuse.

ChatGPT API

ChatGPT’s API includes gpt-3.5-turbo, which is the same as the one employed in the ChatGPT. It comes at a rate of $0.002 per 1k tokens, which represents a 90% reduction from previous GPT-3.5 models. Besides, this model has demonstrated outstanding performance for various non-chat applications. During the preliminary tests, users transitioned smoothly from text-davinci-003 to gpt-3.5-turbo with only minor tweaks required to their prompts.

OpenAI said that there will be continuous model improvements and the ability to select dedicated capacity for greater control over the models. The new release of gpt-3.5-turbo-0301, which will be supported till June 1, and updates to gpt-3.5-turbo will be disclosed in April.

Check out the guide for ChatGPT here.

Whisper API

Whisper, which was released in September 2022, has been widely acclaimed by developers. However, running it can be a challenging task. To address this issue, OpenAI has launched the large-v2 model. This API provides convenient on-demand access at a price of $0.006 per minute. 

The API is available through two endpoints, namely transcriptions (for transcribing in the source language) or translations (for transcribing into English). The API supports various file formats, including m4a, mp3, mp4, mpeg, mpga, wav, and webm.

Check out the Speech-to-Text guide here.

Built on developer-focused feedback, OpenAI said that it has been reviewing the feedback from its API customers for making improvements in their models and has made several concrete changes.