Chatbot Arena Invites Users to Test New Models in Agent and Battle Modes

Chatbot Arena has invited users to test a newly added model or feature across its "Agent Mode" and "Battle Mode" platforms. Users can interact with the update directly on the platform to evaluate its performance. This update allows the community to benchmark the new model's capabilities in both blind head-to-head comparisons and complex, multi-step autonomous workflows. Crowdsourced evaluations on Chatbot Arena help establish objective rankings for LLM performance. The testing is split between Battle Mode, which focuses on side-by-side anonymous chat comparisons, and Agent Mode, which evaluates the model's ability to autonomously plan and use tools. No specific technical details about the model's identity were disclosed in the initial announcement.