Microsoft announced the integration of the model GPT-4 Turbo on OpenAI in its AI chatbot Copilot, formerly known as Bing Chat. This update promises to understand and answer more complex and detailed questions, supporting windows up to 128 thousand characters of text.
Microsoft Copilot integrates OpenAI’s GPT-4 Turbo model
Il current limit of 5000 characters imposed on Copilot for information requests can be a problem for complex questions that require context. With the implementation of the GPT-4 Turbo model, Microsoft aims to significantly expand this limit, allowing users to take advantage of a contextual window up to 128 mila characters.
As Hardware Upgrade reports, Mikhail Parakhin, head of Microsoft’s web services and advertising division, has confirmed that the primary objective of this update is to extend the contextual window. This leads Copilot with GPT-4 Turbo to surpass GPT-4 in terms of “perceived intelligence”.
Interactions with GPT-4 Turbo should demonstrate greater fluidity, consistency e similarity to human interaction. The upgrade not only promises improvements in perceived intelligence, but also faster response timesthanks to a greater efficiency of the model.
However, Parakhin acknowledged the presence of some technical problems that require attention, mainly in areas such as mathematics and programming. These problems prevent the definition of one precise date for full implementation.
Another upcoming innovation is the extended support per plug-in third-party, currently in the testing phase. This will give new possibilities to Copilot’s artificial intelligence, which will become more flexible.