What will the published weekly Sonnet-hour usage cap on the Claude Code Max 20x (highest-tier) consumer plan be as of the end of May 2027, in hours?

rationale:

Current Landscape & Last Known Figures Anthropic currently publishes no official, fixed "Sonnet-hour" figures for its Max 20x plan. The last published estimates date back to August 2025, when the company stated the Max 20x plan allowed for 240–480 Sonnet hours per week https://www.explainx.ai/blog/anthropic-claude-max-lawsuit-usage-limits-2026 Is Claude Max 20x really 20x? https://www.explainx.ai/blog/claude-usage-limits-2026-timeline-explained. Today, Anthropic's help pages describe limits without specific hour counts, noting instead that Max 20x provides 20 times the Pro plan's per-session allowance alongside an overarching weekly limit What is the Max plan? | Claude Help Center How do usage and length limits work? | Claude Help Center https://support.claude.com/en/articles/11049741-what-is-the-max-plan. Furthermore, the introduction of multiple new models in September and October 2026—including Claude Fable 5.1, Opus 5.5, Sonnet 5.5, and Haiku 5.5—dilutes the relevance of a pure "Sonnet-hour" metric, making it unlikely that Anthropic will voluntarily return to such narrow terminology.

Recent Limit Adjustments Changes to usage limits over the past year have exclusively been communicated as relative percentages rather than absolute hours. While the five-hour session limits were doubled in May 2026, the weekly caps were explicitly left unchanged Higher usage limits and a SpaceX compute deal https://www.explainx.ai/blog/claude-usage-limits-2026-timeline-explained. Following a temporary summer promotion, Anthropic instituted a permanent 25% increase to weekly limits starting September 14, 2026 https://support.claude.com/en/articles/15910845-claude-code-may-august-2026-weekly-limits-promotion https://support.claude.com/en/articles/15910845-claude-code-may-august-2026-weekly-limits-promotion. Because no new hour baseline was provided alongside this update, interpreting the true hour cap requires extrapolating from the obsolete 2025 figures.

Resolution Scenarios and the Definition of a "Cap" With no current numbers explicitly published, resolution hinges on how the historical baseline and subsequent percentage bumps are interpreted. The most natural reading of a "cap" drawn from the historic 240–480 range is the upper bound of 480 hours. However, applying the recent 25% permanent increase to that upper bound results in a cap of 600 hours. If a strict lower bound or midpoint is used instead, the figure lands around 240 or 360 hours, but the permanent increase and the stronger upper-bound interpretation pull the 25th percentile to 400. The resulting distribution is inherently lumpy. The median of 490 hours slightly exceeds the historical upper bound while remaining below a full application of the 25% increase, and the 75th and 90th percentiles (550 and 590 hours) capture scenarios where most or nearly all of that scaling is applied. At the same time, aggressive API price cuts or compute scarcity would likely put downward pressure on consumer usage caps, limiting the weight placed on outcomes at or above 600 hours.

Pressure to Publish & Tail Risks There remains a modest risk that Anthropic publishes precise hour figures before May 2027. The proposed class action lawsuit Kahn v. Anthropic, which alleges that weekly caps were not clearly disclosed https://www.explainx.ai/blog/anthropic-claude-max-lawsuit-usage-limits-2026 Anthropic sued over Claude Max subscription usage limits, could force the company to issue transparent, fixed numbers. If compelled to publish, Anthropic's healthy Q3 2026 margins ($1.70 in revenue per dollar of compute) could support higher explicit limits, though the computational weight of newer, agentic models might suppress them. Set against related expectations for API pricing and overall revenue, aggressive price cuts could reduce the resources available to subsidize consumer usage, while compute scarcity would directly constrain the hours Anthropic can promise. The 10th percentile therefore sits at 260 hours, close to the historical lower bound while allowing for the permanent increase, covering scenarios where pricing pressure or compute costs constrain the disclosed cap.