It’s not just AI companies that are seeing sky-high valuations — companies that evaluate their performance are doing pretty well too.
LMArena, the AI evaluation platform that emerged from UC Berkeley’s research community, announced today it has raised $150 million in Series A funding at a $1.7 billion valuation. The round, led by Felicis and UC Investments, represents nearly a threefold increase from the company’s $600 million valuation in its May 2025 seed round just months ago.
The financing comes as LMArena has rapidly scaled its business since launching commercial evaluations in September. The company reports its annualized consumption run rate has surpassed $30 million, with more than 5 million monthly users across 150 countries generating over 60 million conversations per month.
The Series A round included participation from Andreessen Horowitz, The House Fund, LDVP, Kleiner Perkins, Lightspeed Venture Partners, and Laude Ventures. The rapid valuation growth underscores investor conviction that independent AI evaluation represents critical infrastructure as the technology reaches mainstream adoption.
From Academic Project to Billion-Dollar Company
LMArena, formally known as Arena Intelligence Inc., was founded in 2025 by Anastasios N. Angelopoulos (CEO), Wei-Lin Chiang (CTO), and Ion Stoica (Advisor). The platform originated as Chatbot Arena, an academic project at UC Berkeley designed to benchmark large language models through crowd-sourced, blind pairwise comparisons.
The company’s approach differs fundamentally from traditional AI benchmarking. Rather than relying on static test sets, LMArena collects real-world human preference data by having users compare outputs from different AI models in anonymous side-by-side evaluations. This methodology has made it a trusted resource for understanding how models perform in actual usage scenarios rather than controlled environments.
“Our mission is clear: to measure and advance the frontier of AI for real-world use, ensuring that developers, researchers, enterprises, and everyday users can understand how AI behaves where it matters most,” the company stated in its announcement.
Expanding Beyond Language Models
While LMArena built its reputation on language model evaluation, the platform has rapidly expanded into new domains. The company now evaluates model capability and reliability across text, code, image, video, and search.
Recent product launches include WebDev Arena, introduced in December 2024, which tests models on web development tasks including building applications with HTML, CSS, and JavaScript. In March 2025, the company launched Search Arena to evaluate retrieval-augmented models on diverse tasks ranging from coding to current events.
The platform’s live leaderboard, powered by millions of community comparisons, has become a key reference point for AI labs and enterprises assessing model performance. Major AI companies including OpenAI and Anthropic use LMArena’s services for evaluation.
The Business of AI Evaluation
LMArena operates on a freemium model where the public platform drives community engagement and data collection, while revenue comes from private evaluations, custom arenas, APIs, and analytics for enterprise clients and AI labs.
The company is positioning itself in what analysts estimate is an $800-900 million AI testing market in 2025, projected to grow to $3.8 billion by 2032. The funding announcement emphasized that “this milestone reflects a growing industry consensus: AI cannot scale responsibly without independent, transparent, and continuous evaluation.”
With the new capital, LMArena plans to scale its engineering, research, platform operations, and community initiatives. The company emphasized that the “unprecedented engagement signals a fundamental shift in expectations: the world now demands AI that is measurable, comparable, and accountable.” And the rapid ascent from academic project to billion-dollar company in less than a year illustrates the premium investors are placing on AI infrastructure companies, particularly those addressing critical concerns around transparency, reliability, and accountability in AI systems.