Microsoft and Google have new agents and models with augmented thinking – computers
Openai recently announced that Chatgpt has gained new capabilities Create pictures with context. But Sam Al -Taman was not the only one to reveal smart news. Both a Microsoft as Google also advanced with new solutions and models that highlight the advanced thinking thinking.
On the side of microsoft,The news arrives in the form of two agents for Microsoft 365 Copilot. According to the company, the new Researcher and analyst are able to analyze huge amounts of informationAccess to work data in a safe way, as well as the Internet, to respond to complex requests.
On the one hand, the researcher is performed to help users at the expense of complex investigation tasks. The agent combines an Openai’s deep search form with COPILOT research and advanced coordination capabilities.
The company states that the agent can be used, for example, Developing detailed market strategies; ISubmit new opportunities for products From trend data; Or to Create market analysis reports.
Click the photos to see some examples of the researcher’s use
Microsoft details that the researcher also has the ability to use third -party data to improve answers by integrating external sources such as Salesforce, Servicenow and Conflow Services directly on Copilot.
On the other hand, Analyst “Think as Data World”Microsoft says. APeople use Openai O3-MINI model It has been improved for an advanced data analysis, using its thinking capabilities Current answers that “reflect human analytical thinking”Refer to the company.
See more about the video analyst
OS Researchers and analyst agents will start access to Microsoft 365 Copilot users in April, As part of a new program, it is called Frontier, which allows early access to new chatbot features while they are still under development, The company provides.
Google Gemini 2.5 reveals
Google also bets on artificial intelligence with the possibilities of advanced thinking and It now reveals Gemini 2.5, a model capable of “thinking” before responding to user requestsWhich, according to technology, improves performance and a higher level of accuracy.
“With Gemini 2.5, we have reached a new performance level by combining a major advanced model with an improved exercise”Google offers, adding that in the future you plan to integrate the possibilities of advanced thinking in all its models.
S. Programming performance It was one of the areas where Google focused, and with Gemini 2.5, the company states this MakeIncluding the creation of web applications, code applications for agents, as well as converting and editing code.
In the next video, you can watch an example on How the model uses its thinking possibilities to create a game, Create an executable code from one line of code.
Watch the video
Gemini 2.5 Pro has multimedia capabilities, and at the present time, Google advances with A copy with a context window with one million symbols.
The model is Available through Google Ai Studio and Gemini for users who have advanced plansSoon to the head there. Technology indicates that in the coming weeks, it will advance with paid methods for greater capabilities.
