Cursor has launched Cursor Router, an intelligent model router for Teams and Enterprise customers that selects an AI model before each coding request runs. Available now on desktop, web, iOS, the Cursor CLI, and Cursor’s SDK, it targets frontier-level performance while reducing spending. Cursor says early-access customers cut costs by roughly 30% to 50%, while online A/B tests across millions of requests showed frontier-quality performance at 60% savings.
Introducing Cursor Router, our intelligent model router that selects the right model for the task at hand.
— Cursor (@cursor_ai) July 22, 2026
Router delivers frontier-quality results at 60% lower cost. pic.twitter.com/R0YABowFKg
The system uses a classifier trained on more than 600,000 live requests. It examines the query, context, task complexity, domain, and Cursor’s observations of model behavior. Routine work can go to lower-cost models, interface tasks to models chosen for visual taste, and long-horizon problems to frontier reasoning models. Training and production measurements account for cache misses caused by switching models.
Users select Auto in the model picker and choose Intelligence, Balance, or Cost. Intelligence aims to match the strongest available models, Balance targets frontier-level quality at a lower price, and Cost prioritizes capability while controlling token spending. Administrators can enable the router by team or group, restrict modes and individual models, and set a default.
Cursor reports that Auto Intelligence reached satisfaction near Fable at about 60% lower cost and scored roughly 15% above Opus 4.8 at nearly the same cost. Auto Balance exceeded Opus 4.8 at about 36% lower cost and matched GPT-5.6 Sol satisfaction at a lower spending rate. Reported cost per commit was $6.76 for Intelligence and $4.63 for Balance, compared with $12.69 for Fable 5 and $7.34 for Opus 4.8.
The company based its findings on online A/B tests measuring user responses and keep rate, or how much agent-generated code remained in a codebase. During two weeks of early access, three high-volume enterprise accounts with thousands of users reportedly saved 30% to 50% without a drop in quality.
Cursor routes hundreds of millions of coding requests across models and providers each week. With roughly 60% of its developers choosing one daily model, routine tasks can otherwise run at frontier-model prices. Cursor Router addresses that cost gap alongside its tool-calling system, which loads less common tool descriptions only when needed, while Composer handles routine work and Grok 4.5 expands the pool for harder tasks.
