During its first-ever developer conference, OpenAI has launched a suite of new models and developer tools, alongside notable reductions in pricing.
Key highlights from DevDay include the introduction of GPT-4 Turbo, featuring a 128K context window and more affordable rates, the assistants API, and advanced multimodal capabilities encompassing vision, image creation and text-to-speech functions.
The latest iteration, GPT-4 Turbo, surpasses its predecessor with the knowledge updated to April 2023 and an increased context window allowing for more extensive interactions
Significantly, its cost-effectiveness has been improved, making it three times less expensive for input tokens and twice as affordable for output tokens.
We're rolling out new features and improvements that developers have been asking for:1. Our new model GPT-4 Turbo supports 128K context and has fresher knowledge than GPT-4. Its input and output tokens are respectively 3× and 2× less expensive than GPT-4. It’s available now to…
— OpenAI (@OpenAI) November 6, 2023
Reproducible outputs
A new seed parameter has been incorporated to facilitate reproducible outputs, ideal for debugging and unit testing applications.
Meanwhile, an updated version of GPT-3.5 Turbo now supports a default 16K context window and showcases a 38% improvement in format following tasks.
The new Assistants API is poised to revolutionise app development by facilitating the creation of agent-like experiences with capabilities such as the code Interpreter and Retrieval.
This API simplifies the integration of complex AI functionalities into various applications.
GPT-4 Turbo and a bunch of other stuff: https://t.co/0gYM9k3mlf— Sam Altman (@sama) November 6, 2023
Expanded modalities
OpenAI has expanded its modalities to include vision in GPT-4 Turbo, which can interpret images, and DALL·E 3, which offers direct integration for image generation.
Additionally, a new text-to-speech API has been launched, providing human-like speech synthesis.
A custom models program is being introduced, offering organisations the opportunity to work closely with OpenAI researchers to develop highly specialized GPT-4 models.
Alongside these advancements, OpenAI has implemented a reduction in prices across its platform.
Developers can begin to experience these updates immediately, with further enhancements and features expected to roll out in the coming weeks.