Input/output modalities and key model metadata from the catalog.
30-day activity plus recent runtime. Text-first models use token volume; other modalities fallback to request activity.
Only benchmarks with comparable results across every selected model are shown.
Switch benchmark type to compare percent and numerical families separately.
Per-1M normalized pricing from observed provider tiers. Blended total uses 90% input + 10% output.
No providers found yet.
API provider availability and subscription plans.
6 plans
Cursor individual plan for daily agent sessions with higher usage than Pro.
Cursor individual plan for agent power users.
Custom Cursor plan for larger organizations that need invoicing, pooled usage, and advanced security.
