Model evaluation
Scenarios
Left model
Right model
Repeat