Arena has reached $100 million in annualized run-rate revenue eight months after launching its commercial AI evaluation service. The company grew from a UC Berkeley research project into a business that sells model-performance analytics to AI laboratories and enterprises.
Arena is best known for its free crowdsourced leaderboard, where users submit a prompt to two anonymous AI models and choose the better response. The platform uses those comparisons to rank models across text, coding, vision, image generation and other tasks.
The company said in its official announcement that its platform has recorded more than 700 million conversations and 82 million votes from over 10 million monthly visitors. Arena previously said its leaderboard was built from more than 10 million user evaluations.
AI Evaluations Drives Commercial Revenue
Arena began generating revenue in September 2025 when it introduced AI Evaluations. The service gives model developers and businesses deeper performance analysis based on feedback collected from Arena’s community.
Co-founder and CEO Anastasios Angelopoulos told TechCrunch that many people still view Arena as an open-source project and do not realize it generates revenue. He clarified that the company charges customers based on consumption, meaning the reported figure is not recurring revenue in the traditional sense.
Arena’s annualized consumption run rate stood at more than $30 million in December 2025, less than four months after the commercial product launched. The figure has since more than tripled.
The company also recently introduced Agent Mode, which evaluates models on complex, multi-step workflows. Arena said the feature has already reached more than five million turns per month and is growing by 10% each week.
Funding Reached $250 Million
Arena raised a $150 million Series A in January 2026 at a post-money valuation of $1.7 billion. Felicis and UC Investments led the round, with participation from Andreessen Horowitz, The House Fund, LDVP, Kleiner Perkins, Lightspeed Venture Partners and Laude Ventures.
The round followed a $100 million seed investment, bringing Arena’s total funding to $250 million. At the time of the Series A, the company said more than five million monthly users across 150 countries were generating over 60 million conversations each month.
Angelopoulos co-founded Arena with CTO Wei-Lin Chiang, another former UC Berkeley postdoctoral researcher. UC Berkeley professor and Databricks co-founder Ion Stoica also helped establish the company after advising the original research project, which incorporated in April 2025.
Arena competes for spending with post-training and human-labeling providers such as Mercor, Surge and Scale AI. Another crowdsourced model-comparison startup, Yupp, shut down in March 2026.
Featured image credits: Magnific.com
For more stories like it, click the +Follow button at the top of this page to follow us.