July 10, 2023 – OpenAI Makes Groundbreaking Announcements In an exciting update from OpenAI, significant developments were unveiled yesterday. OpenAI announced the full release of GPT-3.5 Turbo, DALL-E, and Whisper APIs, designed to assist developers in enhancing the efficiency of model processing. Additionally, OpenAI revealed their ongoing efforts in developing the next iteration of GPT, namely GPT-4, as well as introducing new features to GPT-3.5 Turbo, all set to be launched in the latter half of this year.
OpenAI disclosed that all AI models used in API calls have been seamlessly upgraded to GPT-4 by default, relieving existing users from the need to switch and allowing them to continue their work uninterrupted.
Moreover, OpenAI unveiled the Whisper API, a cutting-edge speech-to-text AI model capable of transcribing user’s speech, videos, and other media into written text.
Furthermore, OpenAI expressed their commitment to continuously improving the Chat Completions API, primarily focusing on optimizing its computational efficiency. They announced their plans to retire the older models utilized by the Completions API in six months’ time, precisely in January 2024.
As of today, developers can still utilize the Completions API; however, OpenAI has clearly marked the older version as “Legacy Completions” in their developer documentation. OpenAI intends to concentrate their resources on the Chat Completions API, discontinuing the public use of the older models employed by the Completions API.
Upon further investigation, it was found that OpenAI anticipates the discontinuation of the older models and the cessation of the Edits API on January 4, 2024. Starting early next year, users of the Edits API and its associated models will be required to transition from the older models (text-davinci-edit-001 or code-davinci-edit-001) to GPT-3.5 Turbo. OpenAI has provided a comprehensive comparison table between the new and legacy models for a seamless transition.
