LMArena lands $1.7B valuation four months after launching its product

LMArena, a startup that originated as a UC Berkeley research project in 2023, announced Tuesday that it has raised a $150 million Series A round at a $1.7 billion post-money valuation. The financing was led by Felicis, with participation from the university’s investment arm, UC Investments.

The company moved quickly after transitioning into a commercial venture, having previously secured a $100 million seed round in May at a $600 million valuation. With this latest raise, LMArena has brought in $250 million in total funding in roughly seven months.

LMArena is widely known for its crowdsourced leaderboards of AI model performance. Through its consumer-facing platform, users enter prompts, which are simultaneously sent to two AI models, and then select the response they believe performed better. According to the company, these comparisons now involve more than 5 million monthly users across 150 countries and generate around 60 million conversations each month, which power its rankings.

The leaderboards evaluate models across categories, including text generation, web development, vision tasks, text-to-image generation, and reasoning. The platform tests multiple versions of major AI models, including offerings from OpenAI, Google, Anthropic, and Grok, as well as models designed for more specialised use cases, such as image generation and advanced reasoning.

The company began as Chatbot Arena, an open research initiative created by UC Berkeley researchers Anastasios Angelopoulos and Wei-Lin Chiang, initially supported by grants and donations. Over time, LMArena’s rankings became closely watched within the AI industry, particularly by model developers tracking their performance against competitors.

As the company began monetising its platform, it partnered with select AI labs — including OpenAI, Google, and Anthropic — to make their flagship models available for community evaluation. In April, a group of rival researchers published a paper alleging that this arrangement allowed certain model providers to influence benchmark results. LMArena has strongly denied those allegations.

In September, the company formally launched its commercial offering, AI Evaluations, which enables enterprises, AI labs, and developers to commission model assessments using LMArena’s global user community. Less than four months after launch, the service reached an annualised “consumption rate” — the company’s term for annual recurring revenue — of $30 million as of December.

That rapid growth, combined with LMArena’s visibility in the AI ecosystem, helped attract strong investor interest for the Series A. Additional participants in the round include Andreessen Horowitz, The House Fund, LDVP, Kleiner Perkins, Lightspeed Venture Partners, and Laude Ventures.