OpenAI introduces GPT-4 Turbo, customizable ChatGPT and more

OpenAI has unveiled new features and models for developers and ChatGPT users during its first-ever developer conference, DevDay 2023. The conference was held in San Francisco and kick-started with a speech for Sam Altman, CEO of OpenAI.

During the conference, OpenAI introduced GPT-4 Turbo, the newest version of their existing AI model ChatGPT, bringing with it access to world events up to April 2023 and accepting much longer inputs than previous versions.

“It has a 128k context window, so it can fit the equivalent of more than 300 pages of text in a single prompt.”

In addition, OpenAI has released the Custom Models program, which offers selected organizations the chance to work with “a dedicated group of OpenAI researchers to train custom GPT-4 to their specific domain.”

The access to the custom models will be exclusive to the organizations and will not be shared with other customers or used to train different models. Also, the data provided for OpenAI’s custom models will not be used in any other context, according to OpenAI.

Furthermore, OpenAI announced the release of Assistants API, which has new capabilities such as Code Interpreter and Retrieval. This assistant “has specific instructions, leverages extra knowledge, and can call models and tools to perform tasks.”

There is also a new modality, DALL-E 3, OpenAI’s text-to-image model that offers different formats, quality options, and resolutions.

Additionally, OpenAI will provide text-to-speech API, Audio API with six preset voices to choose from, and two generative AI model variants.