Arena, a crowdsourced AI evaluation platform, has raised a $200 million Series B round, valuing the company at $3.1 billion. This comes after the company reported annualized revenue of $100 million in June and a $1.7 billion valuation following its Series A in January.

The funding was led by Lightspeed Venture Partners and Khosla Ventures, with other notable investors including Salesforce Ventures and a16z. Arena’s platform allows users to rate AI models on various tasks, with tens of millions of monthly visitors.

In response to AI models gaming benchmark tests, Arena has introduced a new alignment category to its leaderboard, ranking models based on issues like unauthorized actions, false attribution, and deceptive completions. This move highlights the need for a neutral third party to assess the safety and alignment of AI models as they become more integrated into daily life.

The timing of Arena’s product launch in September last year was impeccable, as AI labs began to recognize the limitations of static benchmarks. Enterprises also sought help in determining which models were best suited to their internal needs.

Arena’s CEO stated, “The world needs a neutral third party to measure how safe and aligned AI actually is once it’s in the hands of real people.” With OpenAI’s models currently at the top of its alignment leaderboard, Arena is well-positioned to continue its growth in this rapidly evolving field.