Notion bets big on integrated LLMs, adds GPT-4.1 and Claude 3.7 to platform

The idea of ​​the productivity platform is betting on the large LLMS models that operate more capabilities of new institutions, including the construction of the GPT-4.1 from Openai and Claude 3.7 in the dashboard.

Although both Openai and Anthropic Start BRITY in building productivity features on their chat platforms, this LLMS fetch into a separate service shows the competitiveness of the space.

Metect has announced the new AI tool group for everyone within the idea of ​​the idea today, including the notes of the artificial intelligence meeting, the research in institutions, the research development, and the ability to switch between GPT-4.1 and Claude 3.7 of humans.

One of the new features allows users to chat with LLMS within the idea of ​​the idea and switch between the models. Currently, the idea only supports GPT-4.1 and Claude 3.7. The idea is to reduce the window and the context.

The company said that the first adopters of the new feature include Openai, RAMP, Vercel and Harvey.

https://www.youtube.com/watch?

Mix the form and control

The idea built features with a mix of LLMS from Openai and Claude, and its own models.

Stay away from pure thinking models makes choosing the model interesting to the concept. Technically GPT-4.1 is not a model of logic, while Claude 3.7 is a hybrid model-capable of behaving as regular LLM and thinking model.

Thinking models have a moment, although a lot of caution that these models are still lying at times. However, while thinking models such as Openai’s O3 (and yes, Claude 3.7 Sonnet) take their time to answer, repeatedly on different scenarios, they are not the best for fast thinking tasks and data collection. Many productivity tasks, such as meeting copies or searches for task data, do not need the strength of the thinking model behind it.

Sarah Sachs, Engineering International, told Venturebeat in an e -mail that the company aims to features that did not sacrifice accuracy, safety and privacy, as well as responding to inquiries at Speed ​​Enterprises.

“In order to achieve a low -creation experience, we have flowed models with internal use and comments from trusted laboratories, in order to make artificial intelligence specialized in concept recovery tasks,” Sachs said. “This preparation helps an idea to understand work needs, give related answers, serve customers with the time of the second branch’s transfer, and maintain customer data safe and compatible.”

Sachs said that hosting and building with different models allows users to “choose the option that best suits their needs – whether this is a more tone, better coding possibilities or faster response times.”

Motions of artificial intelligence meeting and more

AI’s idea of ​​work paths and copying meetings for users, especially if they add an idea to their calendar, and can listen to calls.

Users can use the idea of ​​searching for institutions by connecting applications such as Slack, Microsoft Teams, GitHub, Google Drive, SharePoint and Gmail. Sachs said the idea will discuss the organization’s internal documents, databases and connected applications.

In addition to other documents that have been downloaded or searching on the Internet, the Foundation’s search results allow the IDION users to reach the new search mode. You will formulate the documents directly from the idea during “Analyze all your sources – as well as the web – and thinking through the best response.”

The idea also added GPT-4.1 and Claude 3.7 as chat options. Openai indicated that users in the idea of ​​chatting with GPT-4.1 can on the work space and create a direct idea template from the conversation. Sachs said that the company is working to add more models to its chat feature.

GPT-4.1 is now integrated into an idea!

You can ask GPT-4.1 to help in anything, then create a direct idea page from the caravan.

(Refer to the GPT-4.1 Litage has not been accelerated in this video.) pic.twitter.com/Pvrzpupwrf

Edwinarbus May 13, 2025

Participants in business and institutions plans get the idea of ​​immediate access to new features.

Competition with model service providers

Although users can access both Antarbur and Openai on the platform, the idea still has to compete with models.

Deep Openai’s research is welcomed as a game of augmented reservation generation (RAG). Google also has a copy of deep research. Anthostopis can search on the Internet for you.

Not to mention that the idea needs to compete with other platforms that already benefit from artificial intelligence. The meeting space is full of companies tracking, copying, and summarizing and withdrawing visions of calls with artificial intelligence.

However, the large -scale sale point of the index is that it has all these capabilities on one platform. Institutions can use all these different services, but they live outside the chosen productivity platform. The idea said that the presence of all these features in one place, with everyone’s pricing in one, will provide institutions from participating in various platforms.