Ramp introduces Router for multi-model AI API access

Corporate expense management platform Ramp has launched Router, a United States-only AI model routing service that lets users and companies access and switch between large language models through a single API. Ramp says the service is free to use for the remainder of 2026, although customers still pay model inference costs, and it is offering a $26 launch credit.

Router currently provides access to models from OpenAI, Anthropic, DeepSeek, Moonshot, Minimax, Nvidia, xAI and Z.ai. Ramp says it built and used the routing technology internally for its own AI requirements over the past three years before making it available as a product.

Routing choices and operational visibility

The service is designed to route requests using customer preferences rather than requiring teams to change providers manually. One strategy can favour model providers' flex usage tiers. Another can select a model against as many as three benchmarks specified by the customer.

Users can also send only difficult tasks to more expensive models or test models without changing their integration. Router includes a dashboard showing token spend, cost, latency, fallback attempts and other operational details. These controls put model selection and usage monitoring in the same workflow as inference access.

Data retention is a key deployment consideration

Router records model inputs, outputs and tool calls for one year by default under an opt-out data retention policy. Ramp says it removes personally identifiable information before using that content to improve the product. Companies considering the service will therefore need to assess its retention setting alongside the nature of the prompts, outputs and tool calls they plan to process.

The launch places Ramp in a growing market for services that sit between applications and multiple model providers. The category has drawn attention as platforms seek to make testing, cost management and provider switching easier. The proposed transaction involving Stripe's proposed OpenRouter acquisition shows how AI gateways can become strategically important as model access expands.

How Router fits Ramp's product strategy

Ramp already offers AI token usage monitoring and token spend management, making Router an extension of its existing product set. The company also gains a potential entry point for customers that first need model routing and usage controls before adopting broader expense-management tools.

Ramp raised $750 million at a $44 billion valuation in June. It has not disclosed Router's pricing after 2026, and its current model catalogue is smaller than OpenRouter's. For businesses, the practical implication is to evaluate routing features, model costs, performance data and retention policies together before assigning production AI workloads to a gateway.