Smart Model Selection
Spend less on AI models
Let Mind Code choose an approved model for each task. Save on model costs without making every choice yourself.
| Turn | Band | Score | Category |
|---|---|---|---|
| #14 | complex | 0.82 | refactor |
| #15 | trivial | 0.21 | lookup |
| #16 | standard | 0.47 | unit tests |
Match model to task
Use an affordable model for simple work and a more capable one when needed.
Choose when you want
Pin a model for a session, then return to automatic routing whenever you like.
Know what each request used
See the selected model, why it was chosen, its cost, and response time.
Track spend by team
See where model spend goes across people and teams.
Use the right model for each task
Mind Code chooses an approved model based on what the task needs.
- Keep simple work efficientUse cost-effective models for routine requests.
- Step up when neededChoose a more capable model when the task needs it.
- Support involved workAccount for tasks that use tools or change files.
Avoid needless model changes
Keep a model that fits the work. Switch when the task needs something different.
- Keep useful momentumAvoid switching when your current model is already a good fit.
- Preserve useful contextKeep the work and its context together.
- Switch when the work changesMove to another model when the request needs different capabilities.
See what each request costs
Review the selected model, routing reason, response time, and cost for each request, with usage attributed to people and teams.
- Request-level costSee the model charge for each request in your usage history.
- Understand the choiceSee whether a model was selected automatically or pinned by you.
- Keep work movingIf a model is rate-limited, Router can fall back to another approved model.
Balance cost and capability
Set your preference for cost, speed, or deeper reasoning.
Auto Cost
Favor efficient models for routine requests and step up when a task needs more reasoning.
Auto Balance
Balance speed and capability across everyday coding tasks.
Auto Intelligence
Favor stronger models when a task calls for deeper reasoning.
Control spend without choosing every model
Choose a preference and see how each request is handled.
- 01
Understand the task
Mind Code considers what the request involves.
- 02
Match capability to need
Use an approved model suited to the task and your preference.
- 03
See the cost
Review which model handled the request and its usage.