> run_evaluation (your task, your rubric, your endpoint + 2 frontier models)
your endpoint (grok-4.7): quality 70 | cost $0.00299
claude-sonnet-4-5: quality 60 | cost $0.00154
gpt-4.1: quality 25 | cost $0.00053
winner by quality: your endpoint (70) | winner by efficiency: gpt-4.1