A cheaper model in a good harness beats a frontier model in a generic one. On top of that, deterministic orchestration gives us confidence we actually covered the whole codebase, when you just ask Codex for vulns, you can never be sure it looked everywhere.
Sacada muito interessante. Aqui fala que se tu tiver usando um dos modelos mais baratos mas com um harness bom, supera um modelo bom com harness ruim