Popular AI leaderboard Arena nearly doubles valuation to $3.1B valuation in 10 months
Arena, the company behind the crowdsourced LMArena leaderboard, raised $200 million in a Series B led by Lightspeed Venture Partners and Khosla Ventures, valuing the startup at $3.1 billion. The funding follows Arena reporting $100 million in annualized run-rate revenue in June and comes about 10 months after a $150 million Series A at a $1.7 billion post-money valuation.

Why It Matters
The round underscores growing investor demand for independent tools that evaluate AI behavior as models become better at gaming static benchmarks. Arena's commercial analytics product and new alignment metrics respond to industry needs for real-world assessments of safety and reliability.
Key Facts
- Funding round: $200 million Series B
- Post-money valuation: $3.1 billion
- Lead investors: Lightspeed Venture Partners and Khosla Ventures
- Other investors: Salesforce Ventures, 01 Advisors, Dell Technologies Capital, Endeavor Catalyst, a16z, Felicis, and others
- Prior round: $150 million Series A in January at $1.7 billion post-money valuation
Arena began in 2023 as a UC Berkeley research project that crowdsourced rankings of AI models and has since grown into a commercial business that claims tens of millions of monthly visitors. On Thursday the company said it closed a $200 million Series B financing led by Lightspeed Venture Partners and Khosla Ventures, bringing its valuation to $3.1 billion. The raise follows a period of rapid revenue growth: Arena reported reaching $100 million in annualized run-rate revenue in June, up from a reported $30 million annualized run rate at the time of its January Series A. That earlier round was $150 million at a $1.7 billion post-money valuation, meaning the company’s value has nearly doubled in roughly 10 months. Arena’s consumer-facing leaderboard remains free and lets people submit prompts and rate which model performs better on tasks or “vibes.” In September of last year the company introduced a commercial offering called AI Evaluations, which provides model labs and enterprises with analytics derived from community feedback. Arena says the timing aligned with broader industry concerns that models were learning to optimize for benchmarks rather than for real-world performance. To address those concerns, Arena added an alignment category to its leaderboard to score models on behaviors such as unauthorized action (taking actions it wasn’t asked to take), false attribution (miscrediting statements to the wrong source), and what the company terms “deceptive completion” (claiming to have completed tasks it did not). Preliminary standings show several OpenAI models leading the alignment rankings, while Anthropic’s Claude Opus 5.5 and Claude Fable appear in sixth and ninth place, respectively.
Keep Reading
Asos confirms breach of customer data after hackers send rogue app notification

New York alleges TikTok gave teens, children a placebo safety feature instead of a real one

Goodfire says its new ‘inside-out’ monitors catch rogue AI agents at a fraction of the cost
