Monday, October 12, 2026

Apps & Consumer

Ramp launches Router, its own AI model-routing service

Ramp launched Router, an AI model-routing service free to use through the end of 2026, letting companies switch between models from providers including OpenAI, Anthropic, and DeepSeek through a single API.

Graphic announcing Router by Ramp with text reading 'Cut inference costs in seconds' against a glowing background
Photo: Router

Corporate expense management platform Ramp is following Stripe into AI-inference toll-taking. Ramp launched its own AI model routing service, called Router, on Wednesday evening, letting users and companies switch between large language models through an API; the company says it has used the router internally for its own AI needs over the past three years. The service is only available in the United States for now, is free to use for the remainder of 2026 (users still pay for model inference costs), and comes with a $26 credit launch offer; Ramp didn’t say how much it will cost in 2027.

Router functions similarly to OpenRouter, though OpenRouter offers more model options than Ramp’s current lineup. Router provides access to models from OpenAI, Anthropic, DeepSeek, Moonshot, Minimax, Nvidia, xAI, and Z.ai, along with several “strategies” for routing requests based on customer preferences — including a preference for providers’ flex usage tiers, routing based on up to three user-specified benchmarks, sending only difficult problems to expensive models, or testing models without switching manually. Users get a dashboard showing token spend, cost, latency, fallback attempts, and other details. Router has an opt-out data retention policy: it records model inputs, outputs, and tool calls for one year by default, though Ramp says it removes personally identifiable information before using that content to improve the product.

For Ramp, entering model routing offers a two-pronged opportunity: tapping the AI inference market while offering existing clients a routing service that fits alongside its AI token-usage-monitoring and spend-management products. If Router proves as attractive a model-testing arena as OpenRouter, Ramp may also build long-standing relationships with AI labs and inference providers worldwide — potentially helping the company, which raised $750 million at a $44 billion valuation in June, gain new customers and a new entry point for selling its expense-management products.

Why it matters

Ramp joining Stripe, Databricks, and Rippling in building AI-spend infrastructure shows expense-management platforms racing to sit between companies and their AI bills, wagering that owning the routing layer creates a stickier foothold than usage dashboards alone.