US AI evaluation platform Arena reaches $100 million annual revenue

en.Wedoany.com Reported - Artificial intelligence model evaluation platform Arena has reached an annualized run-rate revenue of $100 million, eight months after launching its commercial services. The platform originated from a 2023 research project at UC Berkeley and is known for its crowdsourced AI model performance leaderboard.

Arena's leaderboard is based on over 10 million user evaluations. Its consumer website allows users to input prompts, send requests to two models, and then judge which model performs better.

While Arena's leaderboard is free for the public, the company began monetizing in September through AI evaluation services, providing model labs and enterprises with in-depth performance analysis from its community.

Anastasios Angelopoulos, co-founder and CEO of Arena, said many people still view the project as an open-source initiative and are unaware that its business has become profitable. Although the company refers to its revenue milestone as ARR, Angelopoulos clarified that the company charges customers based on consumption, so the revenue is not recurring.

In terms of competition, another crowdsourced AI model selection startup, Yupp, shut down in March. Angelopoulos said the company primarily competes for the same market as human annotation startups such as Mercor, Surge, and Scale AI, all of which help model makers optimize their AI after training.

Arena's revenue has grown rapidly as demand from AI providers for post-training optimization services continues to rise. The company announced in January that it had completed a $150 million Series A funding round at a post-money valuation of $1.7 billion, with an annualized revenue of $30 million. According to a report by The Information in April, Handshake's annualized gross revenue from AI training rose from $550 million in January to nearly $1 billion. Mercor's annualized revenue also surpassed $1 billion earlier this year, up from $500 million in September last year.

Arena ranks models across various tasks including text, coding, vision, and image generation, and handles complex long-running workflows through its recently launched Agent mode.

Arena is led by Wei-Lin Chiang, a postdoctoral researcher at UC Berkeley, who serves as Chief Technology Officer, and was co-founded by renowned professor and Databricks co-founder Ion Stoica. The project was incorporated as a company in April 2025. The company has raised a total of $250 million from investors including Felicis, Andreessen Horowitz, The House Fund, LDVP, Kleiner Perkins, Lightspeed Venture Partners, Laude Ventures, and UC Investments.