Meta Shuts Down Internal AI Leaderboard as 'Tokenmaxxing' Trend Sparks Million Dollar Usage Costs

Meta's internal AI leaderboard goes dark after revealing 60 trillion tokens were used in one month, with individual costs reaching $1.4 million.

By: AXL Media

Published: Apr 11, 2026, 8:07 AM EDT

Source: Information for this report was sourced from Fortune

Gamification of Artificial Intelligence Fuels Token Consumption

A Meta software engineer recently introduced "Claudeonomics," an internal leaderboard designed to track AI token usage among the company’s 85,000 employees. The platform awarded staff with titles like "Token Legend" and "Cache Wizard," fostering a culture of "tokenmaxxing" where high consumption is viewed as a primary performance metric. This gamification effort led to a staggering 60 trillion tokens being processed within a single month across the organization. While meant to be a fun engagement tool, the dashboard highlighted the extreme scale at which Meta staff are integrating third party and internal AI models into their daily engineering and administrative workflows.

Financial Impact of High Volume AI Interaction

The cost of this unprecedented token consumption has raised significant fiscal concerns, with a single top ranked user averaging 281 billion tokens in 30 days. Based on the pricing of Claude Opus 4.6, such usage could equate to more than $1.4 million for one individual's activity. Meta CTO Andrew Bosworth has defended these expenses, suggesting that high token spenders are often five to ten times more productive than their peers. Bosworth characterized the investment as "easy money," signaling that the company currently places no upper limit on the resources engineers can deploy to amplify their output through automated agents.

Executive Absence from the Productivity Leaderboard

Despite Meta’s strategic pivot toward a future dominated by artificial intelligence, the company’s highest ranking executives are notably absent from the "Claudeonomics" top 250 list. Neither CEO Mark Zuckerberg nor CTO Andrew Bosworth appeared in the rankings, suggesting that the most intensive use of AI tools is concentrated among specialized software engineers. This disparity reveals a gap between the strategic vision provided by leadership and the practical, high volume implementation occurring at the ground level. While Zuckerberg champions AI impact as a core expectation for 2026, the data suggests he is not personally engaging with the tools at the same frequency as his technical staff.