Claude Code Usage Limit Cost Impact: Net −17% Since Sept 14, 2026

The Short Version (TL;DR)

• Permanent +25% over the pre-promotion baseline — since Sept 14, 2026, standard weekly Claude Code limits are 25% higher than the limit that applied before May 13, 2026, for Pro, Max, Team, and seat-based Enterprise plans [1][2][5].
• The +50% promotion has ended — it ran May 13 through September 13, 2026 and closed at 11:59 PM PT; the permanent limit replaced it the next day [1][10].
• Net −16.7% (≈ −17%) against the promotion — canonical ratio 100 → 150 → 125; Anthropic confirmed the 17% figure itself [1][3][4][9].
• Both headlines are true — 125 ÷ 100 = +25% (raised vs the old baseline); 125 ÷ 150 = −16.7% (cut vs the allowance that just ended). Same limit, different denominator [6][8].
• Same workload = 120% of the standard cap — a team that used all of the promotional cap needs 20% more headroom today (150/125 = 1.2) [4][6].
• Weekly limits only — the promotion did not affect 5-hour session limits, and Anthropic has announced no plan-price or API-billing change [1][6].
• No absolute numbers — Anthropic does not publish weekly limits as fixed message or token counts; every figure above is a ratio [1][6][9].

What changed: a raise against one baseline, a cut against the other

Anthropic framed the change as a raise, and against the pre-promotion baseline it is one: the permanent standard limit is 25% higher than the limit that applied before May 13, 2026 [1][2][5]. The catch is the denominator. From May 13, 2026 the promotion ran at 50% above that baseline and was extended repeatedly through the summer, so users were living at 150 rather than 100 — the promotional allowance, not the baseline [8][10]. The comparison that matters for an existing plan is therefore promotional level to permanent limit, and that comparison is a cut.

Anthropic's own follow-up made the net number explicit: "Compared to today, this works out to a 17% reduction in weekly limits on Claude Code." [3] For planning purposes, take the ratio: weekly Claude Code capacity is now 83.3% of what it was during the promotion, and 125% of the pre-promotion baseline. Neither number is a published hour count — Anthropic publishes none for any plan — so the index is the only honest way to compare the three regimes.

Baseline, promotion, and now: the worked comparison

The cleanest way to see the impact is the canonical ratio — normalize the pre-promotion weekly limit to 100 units:

Example A — team at 100% of the promotional cap (the binding case)

A 5-seat Max team was consuming 100% of the promotional weekly allowance through Sept 13. Today the same workload consumes 120% of the standard cap — the team is over its limit and needs 20% more headroom: one additional seat (5 → 6, i.e. +20%), usage credits, or a deliberate workload reduction [4][6]. If seats cost $100/mo, holding the promotional capacity means budgeting ≈ $600/mo instead of $500/mo — a $100/mo (20%) increase on the seat line.

Example B — team at 60% of the promotional cap (the safe case)

A team that was using 60% of the promotional cap is at 72% of the standard cap today (60 × 1.2). It still has ~28% headroom, so the change does not bind that workload — but growth planning should price against 125, not 150. The 83.3% ratio applies to everyone; the pain shows up at different usage levels.

Example C — cost per unit of capacity

Because the cap falls 16.7% while the seat price stays the same, each paid seat now delivers ~17% less weekly capacity than it did during the promotion. The dollar cost of one unit of weekly capacity rises by the matching 20% ratio (150/125 = 1.2). That is the number to put in front of a client before quoting a fixed-fee retainer that assumes promotional throughput.

The math, once (from the verified research brief)

• Permanent limit vs the promotion: 125/150 = 0.8333, i.e. −16.7% — Anthropic's own rounded "17%" [1][3].
• Permanent limit vs the baseline: 125/100 = 1.25, i.e. +25% — the vendor's own headline [1][2].
• Baseline reconstruction: pre-promotion baseline = promotional cap ÷ 1.5; permanent limit = baseline × 1.25 [1].
• Same workload today: workload at 100% of the promotional cap = 120% of the standard cap (150/125 = 1.2) [4][6].
• Reversing it would take +20%: returning 125 to 150 requires a 25 ÷ 125 = +20% increase, not the 25% in the announcement — another place the denominators matter [9].
• Cost to hold capacity: ≈ 1.2× the seat line you needed during the promotion (e.g., a Max 5x seat at $100/mo ≈ $120/mo equivalent today) — arithmetic illustration, not a published price.
• Ratio-only: no absolute message or token counts exist in any fetched source; six Anthropic surfaces were checked and none publishes a per-plan weekly figure [1][6][9].

What does NOT change

• 5-hour session limits — Anthropic's promotion page states the boost "did not affect" 5-hour limits, and no five-hour change has been announced since [1][7].
• Plan pricing — no price change was announced with the limits change [1][6].
• API / consumption billing — unchanged; consumption-based Enterprise seats have no fixed weekly limits [1][6].
• Claude chat and Claude Cowork — the 50% increase applied to Claude Code only; usage limits for Claude (web, desktop, mobile) and Claude Cowork were unchanged [1].
• Named plans only — the permanent sentence names Pro, Max, Team, and seat-based Enterprise plans, and does not restate the promotion's exclusions, so this page makes no claim about Free or consumption-based Enterprise seats [1].
• No grandfathering, no regional carve-outs — none is mentioned in the vendor documentation or the reporting [1][6].

What agencies should do now

1. Re-plan headcount and credits against 125, not 150. Any growth plan priced on promotional throughput is ~20% short on capacity today [4][6].
2. Re-quote fixed-fee retainers that assumed promotional throughput. If a delivery commitment was priced on the boosted cap, the standard cap needs a capacity buffer or a usage-credit line [4][6].
3. Model your actual usage. Run the Claude Code Usage Limit Cost Impact Estimator on the calculator homepage with your own usage %, seat price, and headcount — it shows the standard-cap percentage, the capacity change, and the dollar cost of holding the promotional capacity.
4. Watch the primary source, not the press. Anthropic's Help Center article is now the authority for this change and carries the vendor's own update stamp; treat any third-party guide published while the promotion was running as superseded [1][8].
5. Check your own usage bars. Anthropic's best-practices article documents the session and weekly usage bars under Settings > Usage in Claude Code, which is the only place your actual limits are visible — the thresholds themselves are unpublished [11].

Frequently asked questions

Sources & Methodology

Everything on this page is ratio-based, and this version of it was rebuilt on September 14, 2026 against the verified research brief for the change (kanban t_19294d02). The primary source is Anthropic's own Help Center article, read in full and machine-verified: it carries an update stamp of 2026-09-14 06:59 UTC and states both the promotion window (May 13 through September 13, 2026, 11:59 PM PT) and the permanent +25%-over-the-pre-promotion-baseline level effective September 14, 2026 [1]. Both @ClaudeDevs posts of August 29, 2026 were read first-hand through X's public embed payload: the announcement of the permanent 25% and the vendor's own "17% reduction" reply [2][3]. The canonical 100 → 150 → 125 index is reproduced identically by the press cluster [4][5][6][9][10]. No absolute weekly message or token counts exist in any fetched source, so this page intentionally stays ratio-based; the earlier single-source Max multipliers (7.5x → 6.25x, 30x → 25x) remain deliberately unused.

• [1] Anthropic Help Center — "Claude Code May–August 2026 weekly limits promotion" (updated 2026-09-14 06:59 UTC): support.claude.com/en/articles/15910845
• [2] @ClaudeDevs (Anthropic) — "Starting September 14, we're permanently raising standard weekly limits in Claude Code by 25% for Pro, Max, Team, and seat-based Enterprise plans" (Aug 29, 2026): x.com/ClaudeDevs
• [3] @ClaudeDevs (Anthropic) — "Compared to today, this works out to a 17% reduction in weekly limits on Claude Code" (Aug 29, 2026): x.com/ClaudeDevs
• [4] BleepingComputer — "Anthropic is cutting Claude Code's current weekly limits by 17%": bleepingcomputer.com
• [5] MacObserver — "Anthropic raises Claude Code weekly limits by 25% for paid plans": macobserver.com
• [6] Notebookcheck — "Anthropic announces a 25% increase to Claude Code limits — but there's a 17% catch": notebookcheck.net
• [7] usingClaude — "Claude Code permanently raises weekly limits by 25%": usingclaude.com
• [8] AICatchup — "Claude Code Limits 25% permanent starting Sept 14" (updated 2026-09-14): aicatchup.com
• [9] Lapaas Voice — "Claude Code Limits Rise 25% From Sept 14": lapaasvoice.com
• [10] Startup Fortune — "Usage Limit Boost Ends Sept 14 with a Real Cut": startupfortune.com
• [11] Anthropic Help Center — "Usage limit best practices": support.claude.com/en/articles/9797557

Accuracy note: the +25% permanent level, the +50% promotion window (May 13 – Sept 13, 2026 at 11:59 PM PT), the −16.7% net ratio and the plan eligibility (Pro, Max, Team, seat-based Enterprise) are taken from Anthropic's own Help Center article and the two @ClaudeDevs posts of August 29, 2026, all read first-hand and verified 2026-09-14. The promotion's exclusions (Free plans, consumption-based Enterprise seats) applied to the promotion only; the permanent level does not restate them, so no claim is made here about those two plan groups. Anthropic publishes no fixed weekly counts, so this page uses ratios only. Dollar examples (e.g., $100/mo seat → $120/mo equivalent) are arithmetic illustrations of the 150/125 = 1.2 ratio, not published prices.