Clawgate Clawgate
Smart model routing

Pay top model prices only when the work needs it.

Your coding agent fires a lot of routine requests that do not need your most expensive model. Smart routing reads each request, judges how demanding it is, and serves the best allowed model for it — so easy work runs on a lighter, cheaper model and your top model is saved for the hard parts. One switch, no tuning.

How it works

The right model for every request, automatically

You stay in control of which models your team may use. Clawgate picks the best fit for each request from that list.

STEP 1

Allow a set of models

Pick the models your team may use, from a strong flagship down to lighter, cheaper options. Leave the forced model off, then flip on smart routing.

STEP 2

Clawgate sizes up each request

It reads each request and judges how demanding it is — a quick file lookup versus a tricky bug fix — then matches it to the best allowed model for that level of work.

STEP 3

The best fit is served

Routine work runs on a lighter, cheaper model. Hard work still gets your strongest model. Nothing changes for your developers.

There is nothing to tune. Clawgate judges each request and balances accuracy against cost for you, so you get the savings without setting a single dial.

Where the savings come from

Stop paying top rates for routine work

Coding agents send a steady stream of small, routine requests: quick edits, file lookups, short questions. Running all of that on your most expensive model is money left on the table. Smart routing sends the easy requests to a lighter model at a fraction of the price, and keeps your top model for the parts that truly need it.

  • Lighter models cost a fraction of a flagship model. Routing sends the easy work there.
  • Hard requests are never forced to downgrade, so accuracy holds.
  • Works alongside your budgets and caps, so the savings stack with your other controls.
  • Automatic on every request. No weights, no config, no babysitting.

What a batch of requests costs you

Without smart routing Top rate on everything

Every request is billed at your top model's rate, even the routine ones.

With smart routing You pay less

Hard requests keep your top model. Routine requests run on a lighter model at a fraction of the cost.

Illustrative. Actual savings depend on your model mix and how much of your work is routine. The more routine traffic you have, the more you save.

What you control

Simple to turn on, safe to leave on

One switch per policy

Turn smart routing on for a team or a project and you are done. There are no weights to set and no formulas to learn.

Your models, your list

Routing only ever picks from the models you allow. It never reaches for something you did not approve.

Nothing changes for devs

Claude Code, Codex, and OpenCode keep working exactly as before. The savings happen quietly behind the scenes.

Priced with no surprises

Sizing up a request uses a small, fast model call. It is passed through at cost — never marked up — and shows on your dashboard as its own line, so you always see exactly what routing costs against what it saves.

Smart routing is one of the ways Clawgate lowers your bill. It is more than a usage meter: it controls spend, cuts token usage, and flags abuse, all from one gateway.

Serve the right model. Pay the right price.

Turn on smart routing and stop overpaying for routine work.