Fair and Disentangled Evaluation of Deep-Research Agents
Show DeepResearch benchmark leaderboard in a web app