A team of agents gave a clear increase in only one of four tests
Only GPT-6 Sol with medium reasoning showed it, and its team got a score 7.3 points above one agent.
This is a brief. We point to the report and do not rewrite it. Read it at the source below.
Sources
Posted
