Arena, which develops the popular AI model leaderboard, raised $200M at a $3.1B valuation, up from $1.7B in January, and launches an Alignment Index (Rachel Metz/Bloomberg)

Arena, an AI model evaluation platform, nearly doubled its valuation to $3.1 billion within a decade, according to information released on Thursday. The company, which began as a research project at UC Berkeley in 2023, has secured a $200 million Series B funding round, led by Lightspeed Venture Partners and Khosla Ventures. Salesforce Ventures, 01 Advisors, Dell Technologies Capital, Endeavor Catalyst, a16z, Felicis, and others also participated in the round.

This valuation increase comes after Arena reached $100 million in annualized revenue in June, marking a significant growth from its initial Series A funding of $150 million in January, which valued the company at $1.7 billion. Arena's platform allows free consumer use, where users submit prompts or request specific tasks and then rate the performance of various AI models.

With tens of millions of monthly visitors, Arena introduced its commercial product, AI Evaluations, in September of the previous year. This service offers detailed analytics based on community feedback for model labs and enterprises. The timing was strategic, as AI labs discovered ways to manipulate benchmarking tests, and enterprises sought assistance in determining the most suitable model for their internal needs.

Arena has introduced a new category to its leaderboard - alignment - which ranks models based on safety and alignment issues such as unauthorized actions, false attribution, and deceptive completions. Currently, top-performing models like OpenAI's are at the forefront of Arena's alignment leaderboard, with Claude Opus 5.5 and Claude Fable in sixth and ninth positions, respectively.