Run a 20-minute multi-model spend audit after Stripe buys OpenRouter

Stripe's deal makes the fee sheet worth checking

Your OpenRouter dashboard probably shows a monthly total that feels small enough to ignore. Stripe announced on 19 August 2026 that it had agreed to acquire OpenRouter. New York Times sources reported the deal at $7.5 billion, although Stripe did not officially disclose a price. In May 2026, OpenRouter processed 100 trillion tokens monthly, serving 8 million users and more than 400 models. Before the deal, OpenRouter had a $1.3 billion post-money valuation after its 26 May 2026 Series B, which was led by Alphabet's CapitalG and raised $113 million. OpenRouter promised that its product, mission, and current commitments would remain unchanged after the acquisition. The deal is a reason to check the bill, not a reason to rebuild the project. For a weekend builder, the scale is background. The useful question is whether your current setup still saves money.

The fee sheet is the part of a model router that can move your cost. A multi-model project is the right place for this audit, because one cheap model can hide a fee on another expensive model. The standard is to keep costs near zero without over-optimizing. The audit below keeps the check short: identify routed spend, check the bring-your-own-key threshold, compare the load fee with direct pricing, and note expense-management options.

The fee sheet, the threshold, and direct pricing decide your bill

OpenRouter's fee table allows customers to use their own keys without a fee for up to $25,000 of monthly routed spend. Staying below that line makes the dashboard mostly convenience; crossing it makes the fee question real. OpenRouter's fee page lists a 5.5% fee with a $0.80 minimum for card credit loads, a 5% fee for crypto loads, and no markup on inference, the model's answer. The charge can appear when you top up, not when a model returns a response.

A router is worth paying for when the fee is smaller than the time you would spend juggling provider accounts, keys, and invoices. A good comparison uses the same usage on both sides. Short prompts and short answers should not be compared against a hypothetical heavy workload, and long generations should not be compared against a tiny test call. The router is worth it when the saved setup time is larger than the fee you would pay. If you use several models, do the comparison for your heaviest models first, then decide whether the rest are worth the same setup.

The bring-your-own-key path is the easiest place to get a clean answer. Keys that point directly to model providers keep the router a convenience layer when routed spend stays below the fee-free line. Loading credits instead changes the math: a small top-up often makes the minimum matter, while a larger top-up less often makes the percentage matter more. The audit should use the top-up pattern you actually follow.

Expense management is the quiet part of the decision. A side project can look cheap in the model dashboard and still be annoying to reconcile. A clear monthly total, a card statement, or a budget alert can make the convenience worth a small fee. Without that, the fee is just another cost.

Run the audit before you switch anything

Each item should be quick, and the outcome should be one of three: keep the router, move to direct billing, or add a second router when one model is cheap direct but the rest are easier to route.

• List your monthly routed spend from the router dashboard and check it against the fee-free bring-your-own-key line.
• Check how you load credits, then note the card or crypto load fee from the fee page.
• Compare the total cost of your top two models against direct provider pricing for the same usage.
• Note whether expense-management options, such as a budget alert or a card, would make the convenience worth the fee.

Near the threshold, the audit may be the only reason to change anything.