Skip to content

Pick your style: Denkbaar

Book a conversation

Articles

History

Cursor 2.0 pairs its own model with multiple agents

·

Cursor introduced Composer and a new interface for working with multiple agents.

A developer assigns tasks across two work lanes on a large screen.
Multiple agents need clear assignments and coherent results. Illustration: Denkbaar.

On 29 October 2025, Cursor introduced version 2.0 and Composer. Composer was the company’s first coding model of its own. A new interface for working with several agents at the same time also appeared. The announcement therefore brought two components together: the model doing coding work and the environment where a user distributes and reviews that work. Source. A screen shows two separate task lanes beside a small computing module.

Its significance lay in organising the work. Multiple agents can form a useful trial only when each has a clear assignment. For a business, dividing tasks therefore becomes a design choice of its own. The relationship between the results needs to be reviewable as well as the code itself. Two developers combine two checked components in one test screen.

A small or medium-sized business can explore this with two small, independent tasks in a test project. Describe the desired behaviour and corresponding check for each task. Have the results reviewed separately and then run together. Track the time spent coordinating and combining them. Compare that with working on the same assignments in sequence. Use the trial to decide whether working in parallel helps with your kind of maintenance. The number of agents is a setting; a useful, coherent result remains the goal.

Sources

  1. Cursor 2.0 and Composer (own model)

Topics: devtoolhistory