Popular AI leaderboard Arena nearly doubles valuation to $3.1B valuation in 10 months

Arena, the AI ranking project that started at UC Berkeley in 2023, has turned its crowdsourced model test into a very pricey business. On Thursday, the company said it raised a $200 million Series B at a $3.1 billion valuation.

That’s a sharp jump from January, when Arena said it had closed a $150 million Series A at a $1.7 billion post-money valuation. In June, it said its annualized run-rate revenue had reached $100 million. At the time of the Series A, it said that figure was $30 million. The company’s own numbers show how fast the market has moved around a service that, until recently, looked like a research side project with a leaderboard attached.

Arena’s core product is still free for consumers. Users submit prompts or ask for vibe-coded projects, then vote on which model handles the task better. The company says it gets tens of millions of monthly visitors, giving it a huge stream of community feedback to feed into its commercial offering.

That commercial product, AI Evaluations, arrived in September of last year. It sells model labs and enterprises detailed performance analytics based on Arena’s crowd data. The timing could hardly have been better. This year, AI labs have been dealing with models that can game benchmarks, while companies have been looking for ways to judge models on their own internal needs instead of leaning on standardized tests.

Arena is now adding a new category to its leaderboard: alignment. The company says that includes problems such as unauthorized action, false attribution, and what it calls deceptive completion. On its preliminary alignment list, a set of OpenAI models sits at the top, with Claude Opus 5.5 and Claude Fable in sixth and ninth place. The Series B was led by Lightspeed Venture Partners and Khosla Ventures, with Salesforce Ventures, 01 Advisors, Dell Technologies Capital, Endeavor Catalyst, a16z, Felicis and others also participating.