Anthropic adjusted Claude Code usage limits effective September 14 and reported a reduction in Claude cache read costs
Pricing change Updated 55% confidence first seen
Coverage indicates Anthropic’s Claude Code usage limits changed starting September 14, including a reduction in the per-period limit and higher weekly limits for certain paid tiers. One roundup also reports Anthropic reduced Claude cache read costs and cut cache-read volume costs significantly.
Anthropic
Claude Code
Claude Fable 5.1
Claude Mythos 5.1
GPT-6 Astra
OpenAI
Astra
Automated Researchers Can Reliably Mitigate Alignment Failures
Ben’s Bites
Cursor
Evaluating LLMs Under Production Parity
Fable 5.1
Hugging Face
Infinite Slop
OpenClaw 2.0
Decision brief
- What changed
- Anthropic-adjusted Claude Code usage limits effective September 14: one coverage source says the promo-period limit decreases to 125 units from 150, while another says standard weekly limits increase by 25% for Pro, Max, Team, and seat-based Enterprise plans. Separate roundup coverage also says Anthropic cut Claude cache-read pricing by 75% to $0.25 per million tokens.
- Why it matters
- For leaders funding AI-assisted software work, this changes the operating envelope for Claude-based developer workflows: teams may face tighter per-period limits in some contexts even as paid plans get higher weekly allowances. If your organization uses Anthropic APIs with prompt caching, the reported cache-read price cut could materially reduce inference costs for repeated-context workloads, which may justify revisiting budget assumptions and usage policies.
- Evidence
- The usage-limit change appears in two independent newsletter-style roundups: Ben’s Bites reports a drop to 125 from 150 units starting September 14, while Zvi reports a 25% weekly-limit increase for Pro, Max, Team, and seat-based Enterprise plans on the same date. Deep Learning Weekly separately reports Anthropic reduced Claude cache-read costs by 75% to $0.25 per million tokens.
- What remains uncertain
- The coverage does not reconcile how the lower 125-unit figure relates to the reported 25% increase in weekly limits, so the exact impact likely varies by plan, time window, or promotional status. The cache-read pricing claim appears in a single roundup and the articles do not specify eligibility, rollout scope, or whether other Claude pricing components changed.
- Monitor next
- Watch for Anthropic’s official pricing and plan documentation clarifying September 14 limit definitions by tier and confirming the cache-read pricing change and its scope.
Analytical support, not advice — assumptions and open questions stated above.