Configure your case
Main task
Priority
Usage scale
OPEN BETA
Model Selector
The selector helps avoid over-specification: choosing the sufficient model for the task, not the largest by default.
Choose the right model level
The idea is not to reward the largest, but to find the type of model that solves the task with criteria, reasonable cost and a more sober footprint.
AISHA recommendation
mini
Mini or flash model
Start with the lightest that already solves the work.
The task "General drafting" does not need to pay for frontier capacity if your current priority is to keep cost and volume under control.
Usually moves near the base text range and better protects cost and consumption when usage scales.
Works well for
- Summaries, classification and stable operational support.
- High volume flows or where you pay per call.
- Processes that are then reviewed by a person or an additional rule.
Avoid using it when
- Long analyses with strong ambiguity.
- Tasks where reliability depends on complex context.
Who it's for
A first useful and actionable reading
People designing AI flows, purchasing providers or needing to justify why a task does not always require the frontier model.
What it will look at
The selector starts from the task, not the provider name
The central question is simple: what level of model do you really need to solve this work reliably.
What it will return
A sober recommendation with visible trade-offs
The tool will not choose the most spectacular model, but the one that best fits the task and context.
Why it matters
Choosing the wrong model is not a technical detail: it changes cost, consumption and product design
This block attacks a specific market inertia: always using more capacity than needed.
Interactive beta available: helps choose the sufficient model type without over-sizing cost and consumption.