Ramp Launches AI Model Routing Service Called Router

Corporate expense management platform Ramp has introduced its own AI model routing service, simply named Router. The tool, which Ramp has used internally for three years, allows developers to access and transition between various large language models through a single API. Currently restricted to users in the United States, Router is free to use for the remainder of 2026. Ramp is also offering a $26 launch credit, though customers must still cover their own underlying AI model inference costs. The company has not yet disclosed what the service will cost starting in 2027.

The new service provides access to models from major AI developers, including OpenAI, Anthropic, DeepSeek, Moonshot, Minimax, Nvidia, xAI, and Z.ai. To help developers manage their workflows, Router offers several routing strategies. Users can configure the system to prioritize flexible usage tiers from model providers or route queries based on up to three custom benchmarks. It also allows teams to direct complex queries to premium models while keeping simpler tasks on cheaper alternatives, simplifying the process of testing new models without rewriting code.

Practitioners can monitor their operations through a dedicated dashboard that tracks token consumption, overall costs, latency, and fallback attempts. However, users should note the platform's data privacy terms. Router features an opt-out data retention policy that saves model inputs, outputs, and tool calls for one year by default, though Ramp states it will strip personally identifiable information before using this data to refine its product.

For Ramp, which secured $750 million in funding at a $44 billion valuation in June, the launch represents a strategic expansion into the AI infrastructure market. By offering token spend management and monitoring alongside model routing, the fintech firm aims to deepen its relationships with enterprise clients. If the tool gains traction, it could also help Ramp establish closer ties with global AI labs and inference providers, creating a new pipeline for its core financial products.