Model Router Surcharges: Calculate the Effective Token Price

Model routers promise to match task difficulty with price and capability, but the invoice may include a platform token rate, credit fee, or blended price. The absolute surcharge is only one number; its percentage impact changes dramatically with the underlying model.

The budgeting mistake is to look at one visible rate and treat it as the whole system. Coding work crosses planning, context retrieval, tool calls, file edits, execution, validation, repair, and review. Every stage can add tokens, runtime, or people. A useful model preserves those layers and then rolls them into cost per accepted result.

What Matters for the Cost Model

• A fixed per-million surcharge is largest in percentage terms on cheap models.
• Routing mix can change with workload difficulty even when total tokens stay flat.
• Cache writes, reads, and outputs may have different fee treatment.
• Retries caused by weak routing belong in the effective price.

These facts change the shape of spend, but they do not determine whether the product is economical. A higher-priced run can be the cheaper outcome if it merges on the first attempt and needs little review. A low-cost run becomes expensive when it wakes repeatedly, occupies compute while idle, or produces a patch that experts must rewrite.

Build the Equation Before the Forecast

Effective price equals provider charge plus platform charge plus expected retry cost. A $0.25 surcharge on a $0.50-per-million blended provider bill raises price 50%. The same surcharge on a $10 provider bill raises it 2.5%. Percentage markup is surcharge divided by the provider charge.

Keep cached input, uncached input, and output separate because their rates can differ by an order of magnitude. Add tool fees, search calls, image or screenshot processing, and regional premiums only when the workflow actually uses them. For runtime, model billing minimums and rounding explicitly. A spreadsheet that assumes perfect per-second billing can understate a block-priced service.

A Worked Monthly Example

Assume 70 million tokens route to a $0.80 blended model and 30 million to a $6 model. Provider cost is $56 plus $180, or $236. A $0.25 surcharge on all 100 million adds $25, producing $261. If routing changes to a 50/50 split, provider cost rises even though the surcharge remains $25.

Run a sensitivity range around the estimate. Use a low, expected, and high case for acceptance rate, output length, retry count, and runtime. Those variables usually move the answer more than a tiny difference in input-token price. The high case should represent a difficult but plausible sprint, not an impossible catastrophe.

Controls That Preserve the Savings

• Export routed model and token category. Make this visible in the run record so the team can connect the control to dollars and accepted output.
• Compute markup by model, not only in aggregate. Make this visible in the run record so the team can connect the control to dollars and accepted output.
• Include retry probability in offline evaluations. Make this visible in the run record so the team can connect the control to dollars and accepted output.
• Alert when the frontier-model share changes materially. Make this visible in the run record so the team can connect the control to dollars and accepted output.

Give every automated run a durable identifier that follows retries, branches, builds, and review. Without that join key, invoices remain disconnected from engineering outcomes. Aggregate by task class, repository, team, model, and trigger so a change in workload mix does not masquerade as a price change.

The Failure Mode to Watch

A router can optimize for a provider-defined objective that differs from yours. Quality labels may not reflect security sensitivity, latency needs, or repository language. Use allowlists and task-specific evaluation rather than trusting one global intelligence setting.

Set a stop condition before the run begins. Repeated failure on the same test, no material repository change, a spend ceiling, or a deadline can all trigger escalation. Stopping is not failure; it protects the budget and gives a human the evidence needed to change the approach instead of financing another identical loop.

How to Decide

Routing is attractive when task diversity is high and telemetry is available. Direct model selection can be cheaper and easier to audit for homogeneous workloads. Run both paths on the same evaluation set and compare accepted-task cost.

Start with a bounded pilot and preserve a control group using the current process. Track model dollars, runtime dollars, tool charges, reviewer minutes, acceptance, cycle time, and post-merge defects. After enough comparable tasks, calculate total cost per accepted result. Use the AI Cost Calculator for the token component, then add the operational layers the API invoice does not show.