At its annual DevDay event, artificial intelligence startup OpenAI unveiled a slew of new capabilities and pricing changes for its AI platform. The enhancements promise to make OpenAI’s technology more powerful, flexible and affordable for developers building real-world applications.
The star of the show was GPT-4 Turbo, an upgraded version of OpenAI’s large language model that can understand and generate human-like text. GPT-4 Turbo has double the context window at 128,000 tokens, allowing it to take in the equivalent of 300 pages of text at once. This expanded memory and reasoning allows more nuanced conversations and complex instructions.
OpenAI says GPT-4 Turbo is also 3x cheaper per token for input and 2x cheaper for output versus the previous GPT-4. For enterprise users, lower pricing means faster payback on AI investments. But it also lowers the barrier for startups and smaller teams to leverage advanced generative AI.
Assistants API unlocks custom AI agents
The launch of Assistants API was another pivotal announcement. This toolset allows developers to build AI agents customized for specific use cases — anything from coding assistants to vacation planners to voice-controlled DJs.
Assistants can leverage capabilities like natural language conversations, executing functions, running code and retrieving external knowledge. The aim of the launch is to unlock a whole new level of intelligence in apps. Assistants can now learn users’ goals and automatically take actions to fulfill them.