Arena Turns From Passive Leaderboard Into Active Agent Runtime With 'Agent Mode'
Arena (via Latent Space)via @arena·medium signal
Arena (formerly LMArena/Chatbot Arena lineage) launched Agent Mode plus Agent Arena, where users run agents on real tasks and feed aggregate metrics — confirmed success, praise vs complaint, steerability, bash recovery, and tool hallucination — into a live leaderboard. It's one of the clearest examples this week of an eval company becoming an execution platform. For builders, it signals a shift toward production-grounded, behavior-level agent metrics over static prompt comparisons.