The new benchmark could soon inspire the development of other open-world environments to train or test AI agents, which also support multi-modal multi-agent interactions. "Currently, the agents in ...
Results that may be inaccessible to you are currently showing.
Hide inaccessible results
Feedback