Run the same analysis across multiple models in parallel and compare results for confidence scoring.
Select exactly 3 to run concurrently.