LMArena, the company behind a crowdsourced leaderboard that ranks AI models by user votes, raised $150m in a Series A round announced on 6 January 2026. The round valued the company at about $1.7bn post-money, according to TechCrunch and Pulse 2.0.
Felicis and UC Investments, the University of California's investment office, co-led the round. Andreessen Horowitz, The House Fund, LDVP, Kleiner Perkins, Lightspeed Venture Partners and Laude Ventures also took part, the company said in a blog post.
The deal came roughly seven months after a $100m seed round in May 2025, which TechCrunch reported valued the company at $600m. LMArena has raised about $250m in total, according to TechCrunch.
“We cannot deploy AI responsibly without knowing how it delivers value to humans.”Anastasios Angelopoulos, Co-founder and CEO, LMArena
What LMArena does
LMArena began in 2023 as Chatbot Arena, an open research project at UC Berkeley built by researchers Anastasios Angelopoulos and Wei-Lin Chiang, TechCrunch reported. It was first funded by grants and donations. Angelopoulos is now the company's co-founder and chief executive.
On the platform, a user submits a prompt that is sent to two anonymous models and then votes for the better answer. The votes feed public leaderboards that rank models on text, web development, vision, image generation and other tasks. Models from OpenAI, Google, Anthropic and xAI are among those tested.
The company said it has collected 50 million votes across its text, vision, web development, search, video and image arenas since its seed announcement, and has run more than 400 new model evaluations. It reported more than 5 million monthly users in 150 countries and about 60 million conversations a month.
A young commercial business
LMArena launched its first commercial product, a paid AI evaluation service for model labs and enterprises, in September 2025. The company said its annualised consumption run rate, its term for recurring revenue, passed $30m in December, less than four months after launch. It has also introduced products including Arena Expert and Code Arena.
The platform has faced criticism. TechCrunch noted that an April 2025 paper by a group of competitors alleged that LMArena's partnerships with OpenAI, Google and Anthropic helped those labs game its benchmarks. LMArena has strongly denied the claim.
Angelopoulos framed the round around measurement. "We cannot deploy AI responsibly without knowing how it delivers value to humans," he said in a statement reported by Pulse 2.0.
What the money is for
LMArena said the new capital will fund platform operations, technical hiring and more research into its evaluation methods. In its own announcement the company said the money would help it build new features and improve the product for its users.
Peter Deng, a general partner at Felicis, said in a statement that progress in AI cannot be measured by lab benchmarks alone. Jagdeep Singh Bachher, chief investment officer of the University of California, said a trustworthy way to measure performance was needed for AI to scale safely. The company did not disclose its headcount.
Company website: lmarena.ai