LMArena, the community platform changing how the world measures AI progress, has raised $150 million in Series A funding. The round values the company at $1.7 billion post-money, nearly triple its valuation following the seed round in May 2025. Felicis and UC Investments (University of California) led the funding, with participation from Andreessen Horowitz, The House Fund, LDVP, Kleiner Perkins, Lightspeed Venture Partners, and Laude Ventures.
The raise reflects a shared conviction across the industry: AI cannot scale responsibly without transparent and continuous third-party evaluations. With billions of people now using AI worldwide, LMArena delivers transparent, real-world evaluation of how frontier models actually perform.
This investment accelerates LMArena's mission to measure and advance the frontier of AI for real-world use. The platform enables developers, researchers, enterprises, and everyday users to understand how models behave where it matters most: in practical, everyday tasks. LMArena will use the funding to operate its platform, expand its technical team, and strengthen its research capabilities.
The platform's community now spans more than 5 million monthly users across 150 countries, who collectively generate more than 60 million conversations every month. This allows for deep analysis of model capabilities in coding, textual reasoning, professional use cases like law or medicine, searching and citing sources, and creative tasks like image or video generation. This extraordinary global engagement underscores a clear shift: the world expects AI to be measurable, comparable, and accountable to real people.
"We cannot deploy AI responsibly without knowing how it delivers value to humans," said Anastasios Angelopoulos, co-founder and CEO of LMArena. "To measure the real utility of AI, we need to put it in the hands of real users. LMArena does exactly this, leveraging feedback from tens of millions of consumers and professionals to set the North Star of the AI industry. Our evaluations use a transparent, open-source methodology to make these insights public for everyone. This funding accelerates the scientific work and community insights that make live evaluation from real users the gold standard for assessing AI in practice."
Peter Deng, general partner at Felicis, emphasized the importance of moving beyond lab benchmarks. "Progress in AI can't be measured in labs by benchmarks alone. It needs to take into account how real people want to use these systems and what they prefer," said Deng, who recently joined LMArena's board as an observer. "We're leading this round because LMArena has built the most trusted, reliable, real-world signal of AI performance. They have become essential infrastructure for every lab and enterprise."
Demand for trustworthy third-party evaluation has surged due to intense competition between AI labs. LMArena works with leading AI labs and enterprises, including OpenAI, Google, and xAI, all drawing on LMArena's evaluations to improve their models for production use cases and user preferences.
The company's strategy is grounded in a core belief: reliable AI requires open standards, methodological rigor, and evidence derived from a diverse panel of real users. LMArena earns revenue by providing paid AI evaluation services to AI labs and enterprises that measure model performance for users across economically valuable industries like software engineering, law, medicine, and scientific research. Its first commercial product launched in September 2025.
LMArena's annualized consumption run rate surpassed $30 million in December, less than four months after launching its AI evaluation product. The company's rapid revenue growth demonstrates strong market demand for rigorous, real-world AI testing.
LMArena began as a graduate research project at UC Berkeley. Wei-Lin Chiang, Anastasios Angelopoulos, and a broader group of more than 20 Berkeley students were focused on how to rigorously evaluate AI models using pairwise human preferences and proper statistical inference, with guidance from their advisors Ion Stoica and Joey Gonzalez.
The work was initially motivated by a very practical need: evaluating Vicuna, one of the early fine-tuned open models, at a time when there were no good ways to assess chatbots using real, in-the-wild user prompts rather than static benchmarks. At the time, the team thought the effort would simply result in a research paper.
What changed was adoption. While many of the contributors were still students, frontier AI models began appearing on the platform before public release, and leading labs started closely following and publicly reacting to the results. Usage grew rapidly into the millions, and the project found itself at the center of how real-world model performance was being understood across the industry. That was the inflection point: realizing that what began as academic research had become critical infrastructure.
Supporting that level of scale, neutrality, and scientific rigor required dedicated engineering, compute, and ongoing platform operations beyond what a university project could sustain, which ultimately led the team to build LMArena as a standalone company.
Continuing the partnership with Felicis was an obvious choice for the founding team. Felicis is known for actively supporting portfolio companies, including recruiting technical, product, and executive talent, customer introductions, and marketing support. Aydin Senkut, Felicis founder, is one of the most active and successful AI investors in the market. Peter Deng brings considerable experience as the former VP of Product at OpenAI who launched ChatGPT and ChatGPT for enterprise.
"Without a trustworthy way to measure performance, AI can't be safely scaled," said Jagdeep Singh Bachher, the University of California's chief investment officer. "LMArena delivers clarity and confidence for researchers, developers, and businesses. As AI adoption accelerates, LMArena's tools are becoming critical infrastructure."
Since announcing its $100 million seed round in May 2025, LMArena has grown far faster than anticipated. In a matter of months, the community has contributed 50 million votes across text, vision, web dev, search, video, and image modalities. The platform has processed 400+ new model evaluations, spanning both open and proprietary models, and collected 145,000 open-source battle data points across text, multimodal, expert, and occupational categories.
These numbers represent real people shaping how AI is measured. The LMArena community has proven that real-world usage can be the backbone for scalable infrastructure to ensure the responsible deployment of AI. The increased competition among AI labs has created a critical need for rigorous, reproducible evaluations. AI labs need actionable feedback on how to improve their models, and enterprises need to know which models perform best for them.
Most of LMArena's investment is focused on engineering, infrastructure, and research, though the company expects priorities to evolve as the platform scales and demand grows. The team is focused on improving the platform for the entire community, and 2026 is shaping up to be an exciting year for both the company and everyone building with it.
