Guide

Reasoning models

Adjust reasoning effort only when the capability is available.

On this page

Inspect accepted effort levels
Set reasoning_effort
Trade-offs

Inspect accepted effort levels

GET /models returns reasoning.effort_levels and reasoning.default_effort for compatible models. If reasoning is absent, do not send the parameter.

Set reasoning_effort

cURL

curl https://code.verboo.ai/router/v1/chat/completions \
  -H "Authorization: Bearer $VERBOO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "<MODEL_ID>",
    "reasoning_effort": "medium",
    "messages": [{"role": "user", "content": "Explain why low latency matters for developer tools."}]
  }'

Effort	When to use
low	Simple, latency-sensitive tasks.
medium	Balanced coding and analysis tasks.
high	Complex problems where quality is the priority.

Trade-offs

Measure on your workload

More effort can increase latency and usage without improving every task. Compare quality, time, and cost with real inputs.