Notion Model Picker Adds Favorites and Adjustable Thinking Effort.

Notion is giving users more control over how they choose and run AI models inside its workspace. A refreshed model picker now lets people star frequently used models and adjust the amount of thinking effort applied to a task.

The company announced the update on LinkedIn, describing it as a facelift for the existing model selection interface. Notion has not published a detailed release note covering availability, eligible plans or whether every model supports each effort setting.

The ability to star models addresses a practical problem created by Notion’s expanding AI lineup. As more options appear in the picker, users need a quicker way to return to the models they rely on most. Favorites should reduce repeated searching and make it easier to maintain different preferences for writing, analysis or other workspace tasks.

This matters more as model choice becomes part of an everyday workflow rather than an occasional technical decision. Notion recently added another option to its lineup, as covered in Notion Adds Grok 4.6 to Its Expanding AI Model Lineup. A favorites layer can keep that growing catalogue manageable without requiring Notion to hide less frequently used models.

For a small business team, starred models could support several recurring patterns:

• Content work: Keep a preferred drafting or rewriting model readily available.
• Research and analysis: Favorite a model selected for more demanding reasoning tasks.
• Routine operations: Return quickly to the model used for summaries, classification or simple edits.

Notion did not say whether favorites are personal to each user or shared across a workspace. The announcement also does not specify whether starred selections synchronize across desktop, web and mobile clients.

The more consequential change is the addition of effort controls. These allow users to dial the AI’s thinking up or down based on the complexity of the request. Notion’s framing is straightforward: not every task needs the maximum amount of reasoning.

That distinction could make the model picker useful beyond simply choosing a model name. A user summarizing a short project update may want a lower effort setting, while someone comparing strategic options or analyzing a complicated document may choose a higher one. The control puts that decision in the interface instead of requiring users to express it indirectly through a prompt.

Notion has not disclosed the exact levels available, how they map to underlying model settings, or whether higher effort changes response time, usage allowances or cost. Users should therefore treat the control as a task-level preference rather than assume a specific technical or billing outcome.

The update also reflects Notion’s broader push to make AI a native layer across its workspace. The company is connecting AI features to pages, meetings and automated workflows, including the ability to share pages directly with custom agents. Its developer platform also supports custom tools for Notion Agents, according to the company’s developer platform page.

For builders and small teams, the refreshed picker is a modest interface change with potentially meaningful workflow benefits. Favorites reduce navigation overhead, while effort controls offer a clearer way to match AI behavior to the job at hand. The remaining questions concern rollout, plan eligibility and how consistently the settings work across Notion’s model catalogue.

What is the Notion model picker?

The Notion model picker is the interface used to select an AI model for a task inside Notion. Its refreshed version lets users star favorite models and adjust thinking effort according to task complexity.

What do Notion’s thinking effort controls do?

The controls let users increase or decrease how much thinking the AI applies to a request. Notion positions lower effort for simpler work and higher effort for tasks that need more reasoning, but it has not documented the precise technical behavior of each setting.

Which Notion plans include the updated model picker?

Notion’s announcement does not identify eligible plans or provide pricing information. It also does not specify a rollout schedule, so users will need to check their own model picker for availability.