Krux
April 4, 2026
Cursor's New AI Can Run Eight Coding Agents at Once
Published: April 4, 2026 at 12:38 AM
Updated: April 4, 2026 at 12:38 AM
100-word summary
Cursor just launched 2.0 with Composer, a coding model that lets you run up to eight AI agents in parallel on a single prompt. Each agent works in its own isolated copy of your codebase, so they won't trip over each other. Most coding turns finish in under 30 seconds, and Composer claims to be 4x faster than comparable models. The pitch: instead of waiting for one AI to finish before starting the next task, you can spin up multiple agents simultaneously. One refactors your API while another writes tests and a third updates documentation. Cursor says this cuts review bottlenecks, though the speed claims come from their own benchmarks,...
What happened
Cursor just launched 2.0 with Composer, a coding model that lets you run up to eight AI agents in parallel on a single prompt. Each agent works in its own isolated copy of your codebase, so they won't trip over each other. Most coding turns finish in under 30 seconds, and Composer claims to be 4x faster than comparable models. The pitch: instead of waiting for one AI to finish before starting the next task, you can spin up multiple agents simultaneously. One refactors your API while another writes tests and a third updates documentation. Cursor says this cuts review bottlenecks, though the speed claims come from their own benchmarks, not independent tests.
Why it matters
The real test is whether your team spends less time babysitting AI edits than they save by running them in parallel.