{"url":"https://www.removepaywall.com/search?url=https://www.theinformation.com/articles/meta-employees-vie-ai-token-legend-status","title":"Meta staff race for Token Legend on AI leaderboard","domain":"removepaywall.com","imageUrl":"https://images.pexels.com/photos/5744248/pexels-photo-5744248.jpeg?auto=compress&cs=tinysrgb&h=650&w=940","pexelsSearchTerm":"social media technology","category":"Tech","language":"en","slug":"0acdbe47","id":"0acdbe47-7e84-4094-af96-c3c793bc94c9","description":"Meta employees compete on an internal leaderboard called Claudeonomics that ranks AI token usage among over 85,000 staff.[[1]](https://www.theinformation.c","summary":"## TL;DR\n- Meta employees compete on an internal leaderboard called Claudeonomics that ranks AI token usage among over 85,000 staff.[[1]](https://www.theinformation.com/articles/meta-employees-vie-ai-token-legend-status)[[2]](https://www.linkedin.com/posts/jyoti-mann-873a4317b_meta-employees-vie-for-ai-token-legend-activity-7446927426221207552-ZdXc)\n- The top 250 users burned more than 60 trillion tokens in 30 days, with the highest averaging 281 billion tokens.[[2]](https://www.linkedin.com/posts/jyoti-mann-873a4317b_meta-employees-vie-for-ai-token-legend-activity-7446927426221207552-ZdXc)[[3]](https://the-decoder.com/meta-employees-compete-for-token-consumption-on-an-internal-ai-leaderboard)\n- This tokenmaxxing trend turns heavy AI compute use into a status symbol tied to perceived productivity at Meta.[[3]](https://the-decoder.com/meta-employees-compete-for-token-consumption-on-an-internal-ai-leaderboard)[[4]](https://x.com/theinformation/status/2041226134348558380)\n\n## The story at a glance\nMeta employees are vying for top spots on a voluntary internal leaderboard named Claudeonomics, which tracks their consumption of AI tokens across various models. The board, set up by one employee on the company intranet, lists the top 250 power users out of about 85,000 tracked staff and awards titles like Session Immortal or Token Legend.[[2]](https://www.linkedin.com/posts/jyoti-mann-873a4317b_meta-employees-vie-for-ai-token-legend-activity-7446927426221207552-ZdXc)[[1]](https://www.theinformation.com/articles/meta-employees-vie-ai-token-legend-status) This is being reported now based on a copy of the dashboard viewed by *The Information*'s Jyoti Mann, amid Meta's broader push to boost internal AI adoption.[[1]](https://www.theinformation.com/articles/meta-employees-vie-ai-token-legend-status) It reflects Silicon Valley's emerging tokenmaxxing culture, where high token burn signals AI fluency.[[3]](https://the-decoder.com/meta-employees-compete-for-token-consumption-on-an-internal-ai-leaderboard)\n\n## Key points\n- Leaderboard called **Claudeonomics**, named after Anthropic's Claude AI, ranks usage and gives badges like Token Legend, Session Immortal, Model Connoisseur, and Cache Wizard.[[1]](https://www.theinformation.com/articles/meta-employees-vie-ai-token-legend-status)[[5]](https://www.techloy.com/at-meta-ai-usage-is-becoming-a-new-performance-metric)\n- Tracks over **85,000** Meta employees, focusing on top **250** heavy users who burn the most tokens via coding, agents, and other AI tasks.[[2]](https://www.linkedin.com/posts/jyoti-mann-873a4317b_meta-employees-vie-for-ai-token-legend-activity-7446927426221207552-ZdXc)\n- In a recent **30-day** period, total token usage topped **60 trillion**; top user averaged **281 billion** tokens, potentially costing millions depending on model pricing.[[2]](https://www.linkedin.com/posts/jyoti-mann-873a4317b_meta-employees-vie-for-ai-token-legend-activity-7446927426221207552-ZdXc)[[3]](https://the-decoder.com/meta-employees-compete-for-token-consumption-on-an-internal-ai-leaderboard)\n- Practice known as **tokenmaxxing**, a form of conspicuous consumption where high usage benchmarks productivity and AI nativeness.[[6]](https://www.linkedin.com/posts/vincekellen_meta-employees-vie-for-ai-token-legend-activity-7446939352317857794-QUqI)\n- Neither CEO Mark Zuckerberg nor CTO Andrew Bosworth ranks in the top 250, per the viewed copy.[[7]](https://www.kucoin.com/news/flash/meta-s-internal-claudeonomics-token-usage-ranking-reveals-60-trillion-tokens-burned-in-30-days)\n- Aligns with Meta's efforts to tie AI use to performance goals, though this leaderboard is unofficial and employee-driven.[[8]](https://www.businessinsider.com/meta-ai-push-employee-goals-tool-adoption-2-026-3)\n\n## Details and context\nAn employee created the Claudeonomics dashboard using internal data to encourage AI adoption, turning token spend into a competitive game. Top ranks signal superuser status, but critics note it may incentivize wasteful usage over efficient output—like past metrics such as lines of code.[[9]](https://www.linkedin.com/pulse/meta-ranking-85000-employees-how-much-ai-use-wrong-metric-shields-7kqxe)\n\nThe **60 trillion** tokens over 30 days could cost around **$900 million** at public rates like Anthropic's Claude Opus ($15/million tokens), though Meta's actual deals and model mix likely lower this.[[10]](https://longbridge.com/news/281816165) The top user's **281 billion** equates to roughly $4 million at those rates, highlighting power-law distribution where a few dominate.[[11]](https://x.com/aakashgupta/status/2041392534296162660)\n\nThis fits Meta's year-long push: access to rival models like OpenAI's GPT and Google's Gemini, plus performance links to AI use. Tokenmaxxing echoes comments from leaders like Nvidia's Jensen Huang, who flags low spend as a red flag for high-paid engineers.[[3]](https://the-decoder.com/meta-employees-compete-for-token-consumption-on-an-internal-ai-leaderboard)\n\n## Key quotes\n> \"Employees at Meta Platforms who want to show off their AI superuser chops are competing on an internal leaderboard for status as a 'Session Immortal'—or, even better, 'Token Legend.'\" (*The Information*, based on dashboard copy viewed by reporter)[[12]](https://www.linkedin.com/posts/markamontgomery_meta-employees-vie-for-ai-token-legend-activity-7447052856723099648-fGAi)\n\n## Why it matters\nHeavy internal AI compute use underscores Meta's aggressive bet on generative tools to boost output amid rising capex.  \nFor employees, token rankings could influence perceptions of value, though they risk gaming over genuine gains; for investors, it flags massive ongoing AI spend.  \nWatch if official metrics formalize this or if efficiency concerns curb tokenmaxxing, especially as costs scale with next models.[[13]](https://tbpn.substack.com/p/four-ai-stories-were-paying-attention)","hashtags":["#meta","#ai","#tokens","#tokenmaxxing","#claudeonomics","#siliconvalley"],"sources":[{"url":"https://www.removepaywall.com/search?url=https://www.theinformation.com/articles/meta-employees-vie-ai-token-legend-status","title":"Original article"},{"url":"https://www.theinformation.com/articles/meta-employees-vie-ai-token-legend-status","title":""},{"url":"https://www.linkedin.com/posts/jyoti-mann-873a4317b_meta-employees-vie-for-ai-token-legend-activity-7446927426221207552-ZdXc","title":""},{"url":"https://the-decoder.com/meta-employees-compete-for-token-consumption-on-an-internal-ai-leaderboard","title":""},{"url":"https://x.com/theinformation/status/2041226134348558380","title":""},{"url":"https://www.techloy.com/at-meta-ai-usage-is-becoming-a-new-performance-metric","title":""},{"url":"https://www.linkedin.com/posts/vincekellen_meta-employees-vie-for-ai-token-legend-activity-7446939352317857794-QUqI","title":""},{"url":"https://www.kucoin.com/news/flash/meta-s-internal-claudeonomics-token-usage-ranking-reveals-60-trillion-tokens-burned-in-30-days","title":""},{"url":"https://www.businessinsider.com/meta-ai-push-employee-goals-tool-adoption-2-026-3","title":""},{"url":"https://www.linkedin.com/pulse/meta-ranking-85000-employees-how-much-ai-use-wrong-metric-shields-7kqxe","title":""}],"viewCount":2,"publishedAt":"2026-04-08T09:28:19.230Z"}