How Cursor Router chooses the right model for the task

Summary

Cursor Router selects models for developer turns using production behavior rather than benchmark scores, balancing user satisfaction and cost. Its routing system first uses Compass, a continuous 0-to-1 complexity predictor trained on follow-up behavior, to keep simple turns on a price-efficient model or upgrade complex ones. Then a taxonomy of domains, tasks, and modifiers compares frontier-model strengths, while an optimizer applies a one-sided 75% uplift threshold and each mode's cost budget. The dataset contains hundreds of thousands of live Cursor turns, preserving conversation signals, model-switching effects, inferred performance from subsequent user actions, and token-based costs including cache misses. Current results report Auto Intelligence at above Fable-level satisfaction with 68% lower cost, and Auto Balance outperforming Opus 4.8 at 41% lower cost while raising satisfaction by 3%.

Takeaways

1. 1

   Compass assigns each turn a continuous complexity score from 0 to 1; its highest-rated turns received a positive performance signal 96% of the time, compared with 71% for the lowest-rated turns.

2. 2

   The routing dataset contains hundreds of thousands of sampled live turns and measures performance from subsequent user actions while calculating cost from API pricing and token usage, including cache misses caused by model switching.

3. 3

   Auto Intelligence delivers above Fable-level user satisfaction at 68% lower cost, while Auto Balance outperforms Opus 4.8 at 41% lower cost and increases user satisfaction by 3%.