Microsoft Reportedly Working on a Smart Mode for Copilot, Could Be Unveiled Alongside OpenAI’s GPT-5

Microsoft is reportedly developing a new feature for Copilot, which could be powered by OpenAI's upcoming GPT-5 model. As per the report, Copilot is getting a new “Smart” mode, which can automatically deduce when to apply reasoning for a query and when to switch to fast responses. OpenAI CEO Sam Altman has also mentioned that the GPT-5 artificial intelligence (AI) model will mark the unification of the GPT-series and the reasoning-focused o-series models. The Redmond-based tech giant's new Copilot feature is reportedly not available to users currently.

Copilot Could Soon Get a Smart Mode

With AI companies releasing multiple models for specific use cases (reasoning, vision, web search, coding, and so on), model switchers have become a commonplace feature in chatbots. It is a dropdown menu users can find on the chatbot's interface, which allows them to select a particular large language model (LLM) for a use case.

However, this dropdown menu has been consistently growing in size, making finding AI models tricky and a cumbersome process. In February, speaking about the roadmap for GPT-5, Altman said, “We hate the model picker as much as you do and want to return to magic unified intelligence.” He also added that he wanted to create an experience where the AI just works without any complications.

Microsoft's Copilot, which is powered by OpenAI's models, is reportedly also working on developing features that apply the same sentiment. According to a TestingCatalog report, the tech giant is planning to add a Smart mode in its chatbot that will alleviate users' confusion about which AI model to choose.

The publication claimed to have found evidence of this Smart mode within the code of Copilot. While it is not currently available for anyone, including beta testers, TestingCatalog was able to activate the feature. Based on a video shared, the Smart mode shows up in the dropdown menu within Copilot's text field. It reportedly joins with existing options such as Quick Response, Think Deeper, and Deep Research.

As per the report, the description of the feature says it “uses GPT-5 to think quickly or deeply,' which also corroborates with what Altman has said about the upcoming frontier model. Since the publication was not able to use the feature, how it exactly functions is currently not known. However, it is expected that Microsoft will release the tool whenever the GPT-4 successor arrives.

Separately, a report claims that OpenAI could release GPT-5 in August, with the expected timeline being the first half of the month. The LLM is said to arrive with reasoning capabilities, as well as access to all the company's tools, such as Canvas, Web Search, Deep Research, and more. GPT-5 is also said to arrive with a mini and a nano variant.