The two sides will jointly build an agent evaluation benchmark for scientific experiment scenarios.
Related