LMArena

【Verdict】LMArena (originally Chatbot Arena) is an open evaluation platform that uses crowdsourced blind battles to measure human preference. Its primary value is for model developers and tool selectors who want to see which model real users actually prefer in side-by-side, anonymous comparisons, with the key output being continuously updated preference leaderboards and task-specific rankings.

【Practical role, threshold & cost】The platform enables zero-cost model comparisons: visit the website, enter a prompt, and two anonymous models respond; you vote for the better one, then their identities are revealed. No login is required for basic usage (optional Google sign-in). The website is free to use, and many community posts describe it as “forever free.” However, this “free” refers only to the randomized blind-battle mode—not guaranteed, stable API access to any specific model. No official pricing exists; all cost impressions come from user demos. It suits general users and product managers who want to quickly feel differences in tone, style, or reasoning among models. It is not suitable for developers who need a consistent API key, reliable latency, and predictable billing.