VentureBeat September 13, 2024
OpenAI surprised the world yesterday afternoon by announcing not “Strawberry” as rumored, nor GPT-5, but a new family of “reasoning” large language models (LLMs) called o1 that aims to offer high performance and accuracy on tasks related to science, technology, engineering and math (STEM) fields.
OpenAI’s two new models are o1-preview and the lower-parameter (less advanced) o1-mini, available now to ChatGPT Plus users as well as developers who use OpenAI’s paid application programming interface (API). This way, developers can test them as the backend of existing third-party apps and services, or build new apps and services atop them.
The new o1 models use a form of “reasoning,” according to OpenAI, and they “try different strategies, recognize mistakes, and are doing...
Today's Sponsors
Today's Sponsor