> For the complete documentation index, see [llms.txt](https://docs.kontinent.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.kontinent.ai/best-practices/cost-quality-tuning.md).

# Configuring the smart router

The smart router only picks among models you put in its pool. Classification decides which request goes where. When that pool mixes cheaper open-weight models with stronger ones, a **cost-quality threshold** from 1 to 10 controls how often the smaller models are allowed to win. There is no code change involved.

## Pick the pool first

The trade-off, if you want one, comes from the models you include — not from Kontinent. A frontier-only pool has no cost/quality dial to speak of. An open-weight plus frontier pool does.

Then set the threshold to match the workload:

| Workload                                                      | Suggested setting |
| ------------------------------------------------------------- | ----------------- |
| High-volume, latency-sensitive (autocomplete, classification) | Low (1 to 3)      |
| General chat and everyday assistant traffic                   | Middle (4 to 6)   |
| Complex reasoning, long-context, or high-stakes output        | High (7 to 10)    |

A low setting leans on smaller and open-weight models wherever they suffice. A high setting favors quality and sends more traffic to top-tier models.

## Combine with sovereignty and custom routers

The threshold controls how the classifier ranks models in your pool; it does not override a model's `sovereignty` tier or a custom router's method. You can set an aggressive threshold for cost while still requiring `strict` sovereignty, or layer a custom router on top for latency- or reliability-based selection. See [Price & performance](/models-and-routing/price-performance.md) and [Maximum availability](/models-and-routing/availability.md).

## Revisit it, don't set it once

Traffic mixes change. If you notice quality complaints on a subset of prompts, raise the threshold or take the weaker models out of the pool; if spend grows faster than traffic, lower it or add open-weight models. Because the setting takes effect immediately with no redeploy, it is cheap to experiment with.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://docs.kontinent.ai/best-practices/cost-quality-tuning.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.