Grok 4.5 achieved a 37.3% Pass@1 and 44.2% mean score in investment banking tasks, securing its fifth-place ranking. In corporate law, the model scored 25.4% Pass@1 and 52.2% mean, indicating it often progresses through legal tasks but struggles with perfect completion.
So What
Grok 4.5 demonstrates specific capabilities in professional services, but its low Pass@1 in corporate law suggests agentic deployment for legal tasks remains challenging.