The Leaderboard “you Can’t Game,” Funded By The Companies It Ranks

Artificial intelligence models are multiplying fast, and title is stiff. With truthful galore players crowding nan space, which 1 will beryllium nan best — and who decides that? Arena, formerly LM Arena, has emerged as nan de facto nationalist leaderboard for frontier LLMs, influencing funding, launches, and PR cycles. In conscionable 7 months, nan startup went from a UC Berkeley PhD investigation task to being weighted at $1.7 billion. 

Watch arsenic Equity host Rebecca Bellan catches up pinch Arena co-founders Anastasios Angelopoulos and Wei-Lin Chiang about really their level became nan go-to leaderboard for frontier AI models, and really they’re trying to build a neutral benchmark moreover arsenic companies for illustration OpenAI, Google, and Anthropic backmost nan project.

They break down really Arena useful and why it’s harder to crippled than fixed benchmarks, what “structural neutrality” really means, why Claude is presently topping master leaderboards successful ineligible and aesculapian usage cases, and really nan institution is expanding beyond chat to benchmark agents, coding, and real-world tasks pinch a caller endeavor product.

Subscribe to Equity on YouTube, Apple Podcasts, Overcast, Spotify and each nan casts. You also can follow Equity on X and Threads, astatine @EquityPod. 

Subscribe for nan industry’s biggest tech news