OpenAI Launches GPT-4 Turbo AI Model with Support for Inputs of Up to 300 Pages in Length

[Tsing]

OpenAI has announced new models and developer products during today's DevDay event, and chief among them appears to be GPT-4 Turbo, a new version of its popular multimodal model that features not only more up-to-date data, but also support for even lengthier content. According to a blog post that OpenAI shared this afternoon, GPT-4 Turbo supersedes the original with a 128k context window, allowing users to paste in content of up to 300 pages in length—a big upgrade over the original's 3,000-ish words. Additionally, GPT-4 Turbo is based on training data that goes up to April 2023, meaning that it can provide the answers to even more important questions, like who won the Super Bowl in February 2022. GPT-3.5 Turbo, another update that improves upon the original with a 16K context window by default, was also announced by OpenAI today.

See full article...

 

[Uvilla]

Been starting to use the bing app, with the AI powered chat, ill admit, for information, that thing has been remarkable so far. The image generator is also remarkable. Above dall e 2 and the other one that does paintings that you use in discord ( forgot the name). These tools are powered by open AI, but I feel MS supped them up somehow. I don't know.