ChatGPT and Claude are ‘becoming capable of tackling real-world missions,’ say scientists
The scientists developed a tool called "AgentBench" to benchmark LLM models as agents.
Search by tag: #missions