{"input_url":"https://developer.nvidia.com/blog/how-to-evaluate-ai-agents-from-tool-calls-to-task-completion/","gnews_decode":"https://developer.nvidia.com/blog/how-to-evaluate-ai-agents-from-tool-calls-to-task-completion/","http_status":200,"final_url":"https://developer.nvidia.com/blog/how-to-evaluate-ai-agents-from-tool-calls-to-task-completion/","html_length":269929,"trafilatura_result":"ok","text_length":12239,"text_preview":"When you ship an AI agent, the key question is whether it can execute a chain of work across dozens of sequential tool calls against a live environment, and recover when a step fails. Scoring whether the model sounds right tells you almost nothing about whether the work finished.\nThat gap is why age","verdict":"OK"}