Ramp has launched its own AI model routing service, called Router, allowing users and businesses to switch between different large language models via an API.

The service is currently only available in the United States and will be free until the end of 2026. Users benefit from a $26 credit launch offer, with cost details yet to be disclosed for future years.

Router’s strategies include setting preferences for model providers’ flexibility tiers, or letting Router decide which model to use based on specific benchmarks provided by users. This could help businesses route difficult problems more effectively without incurring higher costs unnecessarily.

The dashboard offers insights into token spend and latency, among other details, giving users a clear overview of their AI usage patterns. Ramp’s approach also includes an opt-out data retention policy, with personally identifiable information removed before product improvements are made.

For Ramp, this move positions them well in the rapidly expanding AI inference market while integrating seamlessly with their existing expense management products. If Router proves popular, it could help establish long-term relationships with global AI labs and models, potentially boosting Ramp’s customer base and product offerings.