Anthropic doubles Claude Code session cap for paid plans

Coding agents are only useful if they can stay on the job. That has been the annoying gap with Claude Code — developers could get deep into a debugging or refactor session, then hit a usage wall halfway through. Anthropic’s fix is simple and very tangible: as of May 6, it doubled Claude Code’s 5-hour rate limits for paid plans, removed peak-hour reductions for Pro and Max users, and raised API rate limits for Claude Opus too. Anthropic framed all three changes as effective immediately for its most committed users. (anthropic.com) ### What exactly changed? Anthropic made three product changes at once. First, Claude Code’s 5-hour usage limits were doubled for Pro, Max, Team, and seat-based Enterprise plans. Second, Pro and Max users no longer get their Claude Code limits reduced during peak hours. Third, developers using Opus through the API got higher rate limits as well. Ant(anthropic.com)ruptions for heavy users. (anthropic.com) ### Why does the 5-hour cap matter so much? Claude Code is not just chat in a sidebar. It reads a codebase, edits files, runs tests, and works through multi-step tasks in a terminal-style workflow. That means people tend to use it in long bursts — the exact pattern that makes a rolling 5-hour cap painful. If the tool stops in the middle of a real se(anthropic.com)ometimes the willingness to keep using the agent at all. Anthropic itself pitches Claude Code as an agentic coding system built for sustained work across a repo. (anthropic.com) ### Why remove peak-hour throttling too? Because doubling the cap would still feel shaky if the ceiling dropped when traffic spiked. The peak-hours reduction was basically a second limiter layered on top of the main one. Anthropic only removed that reduction for Pro and Max, not every tier, but that still changes the everyday experience for individual pai(anthropic.com)n without wondering whether lunchtime demand will quietly shrink their runway. (anthropic.com) ### What changed for API users? Anthropic also raised rate limits for Claude Opus on the API side. That matters for companies and tool builders running coding agents programmatically, not just inside Claude’s own interface. If Claude Code is the direct user-facing product, Opus API throughput is the plumbing behind custom workflows, internal tools, (anthropic.com)Anthropic did not just tweak the chat app — it widened the pipe for developer infrastructure too. (anthropic.com) ### Why now? Anthropic tied the increase to a new compute deal involving SpaceX and xAI’s Colossus 1 data center in Memphis. The basic story is capacity: more compute gives Anthropic more room to let paying users run longer and heavier workloads. Outside coverage of the announcement described the higher limits as a direct result of that deal, and An(anthropic.com)ent. (anthropic.com) ### Who benefits most? The obvious winners are developers doing long debugging sessions, multi-file refactors, and agent-driven coding loops that can run for hours. Team and Enterprise customers benefit too, because shared work patterns make interruptions even more expensive — one person hitting a wall can stall reviews, fixes, or handoffs. And API (anthropic.com) suggests Anthropic is trying to smooth both the end-user and platform sides of its coding business. (anthropic.com) ### Is this just a quota bump? Not really. It is also a signal about where the product is headed. Anthropic has been pushing Claude as a serious coding and agent platform, and its newer Opus models are explicitly marketed around sustained performance on long-running tasks. Higher limits make that pitch feel more believable. A coding agent that can w(anthropic.com)n is the version people can actually build habits around. (anthropic.com) ### Bottom line This is a small change on paper, but a meaningful one in practice. Anthropic did not launch a new model here — it removed a bottleneck. For people using Claude Code as a real development tool, that may matter almost as much.