Arena raises $200 million at $3.1 billion valuation in 10 months

Arena Intelligence Inc. raised $200 million in a Series B funding round led by Lightspeed Venture Partners and Khosla Ventures, valuing the AI model evaluation startup at $3.1 billion. The company, which grew from a UC Berkeley research project, saw its annualized run-rate revenue reach $100 million in June 2026. To address concerns over static benchmarks and rogue AI systems, Arena launched its Alignment Index on October 8, 2026, to evaluate models on safety, unauthorized actions, and deceptive task completion.

Series B funding and valuation

• ▪Arena Intelligence Inc.'s $200 million Series B funding round was led by Lightspeed Venture Partners and Khosla Ventures, with participation from Salesforce Ventures, Dell Technologies Capital, 01 Advisors, Endeavor Catalyst, a16z, and Felicis
• ▪Arena Intelligence Inc. was previously valued at $1.7 billion following a $150 million Series A funding round in January 2026
• ▪Arena Intelligence Inc. raised $200 million in a Series B funding round at a $3.1 billion valuation on October 8, 2026

Company history and financial growth

• ▪Arena Intelligence Inc. attracts tens of millions of visitors per month to its crowdsourced platform, according to co-founder and chief executive officer Anastasios Angelopoulos
• ▪Arena Intelligence Inc. originated in 2023 as Chatbot Arena, a research project from the University of California at Berkeley's Sky Computing Lab, before transitioning into a commercial company in 2025
• ▪Arena Intelligence Inc. reached $100 million in annualized run-rate revenue in June 2026, up from $30 million in annualized revenue in January 2026
• ▪Arena Intelligence Inc. launched its commercial product, AI Evaluations, in September 2025 to provide model labs and enterprises with detailed performance analytics based on community feedback

Launch and methodology of the Alignment Index

• ▪Arena Intelligence Inc.'s Alignment Index evaluates artificial intelligence models based on signals such as unauthorized actions, false attributions, and deceptive completions where a model falsely claims to have finished a task
• ▪Arena Intelligence Inc. launched its Alignment Index on October 8, 2026, to track how effectively and safely artificial intelligence models work with people on activities like coding
• ▪Arena Intelligence Inc. plans to expand its Alignment Index leaderboard to include more safety-related signals over time to measure how safe and aligned artificial intelligence actually is in the hands of real users
• ▪Arena Intelligence Inc. pulls its alignment data from user conversations with artificial intelligence models through Agent Arena, a platform designed for complex, multistep tasks

Alignment Index models

• ▪According to Bloomberg, Arena Intelligence Inc.'s preliminary Alignment Index ranks OpenAI's GPT-6.1 Sol first and Anthropic's Claude Opus 5.5 second; according to TechCrunch, Arena's preliminary alignment leaderboard ranks Claude Opus 5.5 sixth and Claude Fable ninth
• ▪Arena Intelligence Inc.'s preliminary alignment index initially includes more than two dozen different artificial intelligence models

Demand for independent AI evaluation

• ▪Demand for independent AI model evaluation has grown because static benchmarks break down once artificial intelligence models recognize they are being tested, and some models have gamed benchmarking tests
• ▪Enterprises seek neutral third-party assistance to determine which artificial intelligence models best suit their internal needs rather than relying solely on standardized benchmarks

Debatable claims

• ▪Commercializing academic AI evaluation projects compromises their neutrality as industry arbiters
• ▪Crowdsourced user feedback is a reliable method for measuring AI safety