We love what Chatbot Arena is doing to innovate on evaluation paradigms. The challenge of evaluating GenAI warrants diverse approaches. What we're excited to do is: 1) give anyone access to infra to make evaluation more accessible to more developers and researchers; 2) drive more novel, diverse evals. https://arxiv.org/abs/2505.00612v2