Anthropic tweaks timed usage limits to discourage Claude demand during peak hours

Summary

Anthropic on Wednesday adjusted its opaque usage limits for Claude customers by reducing the power of the services it delivers during times of peak demand, in an effort to balance demand with its capacity to deliver service. In a social media post, Thariq Shihipar, a member of Anthropic's technical team, wrote: "To manage growing demand for Claude we're adjusting our five hour session limits for free/Pro/Max subs during peak hours. Your weekly limits remain unchanged." The change means that during peak hours – 05:00 – 11:00 PT or 13:00 – 19:00 GMT – Claude users could burn five hours’ worth of usage time in under five hours. At other times of day, a five-hour access allowance will allow five hours of access. That odd definition of timed use is possible because Anthropic ties hourly use to token consumption – without revealing exactly how many tokens it ties to timed use. According to Shihipar, "~7 percent of users will hit session limits they wouldn't have before, particularly for pro tiers. If you run token-intensive background jobs, shifting them to off-peak hours will stretch your session limits further." Anthropic has expanded capacity during other times of day when demand is lower, so there's no net loss in terms of usage limits. "Overall weekly limits stay the same, just how they're distributed across the week is changing," Shihipar explained. "I know this was frustrating. We're continuing to invest in scaling efficiently. I'll keep you posted on progress." Anthropic sells its AI services in two forms: an API and subscriptions. API customers pay a published rate for various forms of token usage – Base Input Tokens, 5m Cache Writes, 1h Cache Writes, Cache Hits & Refreshes, and Output Tokens. Subscription customers – Free, Pro ($20/month), Max 5x ($100/month), and Max 20x ($200/month ) – can use Claude subject to unpublished usage limits. Anthropic does not specify exactly how it calculates those limits and users don’t have any way to plan for token usage. "Your...