The leaderboard “you can’t game,” funded by the companies it ranks

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

Watch arsenic Equity host Rebecca Bellan catches up with Arena co-founders Anastasios Angelopoulos and Wei-Lin Chiang about however their level became the go-to leaderboard for frontier AI models, and however they’re trying to physique a neutral benchmark adjacent arsenic companies similar OpenAI, Google, and Anthropic backmost the project.

They interruption down however Arena works and wherefore it’s harder to crippled than static benchmarks, what “structural neutrality” really means, wherefore Claude is presently topping adept leaderboards successful ineligible and aesculapian usage cases, and however the institution is expanding beyond chat to benchmark agents, coding, and real-world tasks with a caller endeavor product.

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

Subscribe for the industry’s biggest tech news