Anthropic is permanently increasing the standard weekly usage limits for Claude Code by 25% for Pro, Max, Team and seat-based Enterprise customers, but the change comes with an important catch for users enjoying the current temporary boost. The company has confirmed that the existing 50% promotional increase will remain available until September 13, after which the new permanent limits will take effect on September 14, 2026.
Because the current allowance is 50% above the original baseline, the new permanent 25% increase will actually represent about a 17% reduction from the amount of Claude Code usage customers have today. Anthropic initially highlighted the permanent increase without emphasizing that comparison, prompting users to point out the difference. The company later acknowledged the 17% reduction and said it is working on changes that will give customers more visibility and control over their usage.
Anthropic Raises Claude Code Weekly Limits By 25%
Anthropic announced that it will permanently raise Claude Code’s standard weekly usage limits by 25% starting September 14.
The change applies to four major paid customer categories:
- Claude Pro
- Claude Max
- Claude Team
- Seat-based Enterprise plans
The company is currently providing users with a temporary 50% increase over the normal weekly limits.
That temporary boost will remain active until September 13 before being replaced by the new permanent allowance.
Claude Code Usage Change
| Usage Level | Relative Allowance |
|---|---|
| Original standard limit | 100% |
| Current temporary boost | 150% |
| New permanent limit | 125% |
| Permanent increase over original | +25% |
| Change from current temporary level | ~17% lower |
The distinction between the original baseline and the current temporary allowance is central to understanding the announcement.
Why A 25% Increase Means A 17% Cut For Current Users
The mathematics behind the announcement is straightforward.
Assume the original weekly Claude Code allowance was 100 units.
Anthropic’s temporary promotion increased that amount to 150 units.
The new permanent increase will take the original 100-unit baseline to 125 units.
That means users will have more capacity than they had before the promotion, but less capacity than they have today.
Claude Code Limit Math
Original limit
100 units
│
▼
Temporary +50% boost
150 units
│
│ September 14
▼
Permanent +25% limit
125 units
The reduction from 150 to 125 units is:
25 ÷ 150 = 16.7%
Rounded, that is approximately a 17% reduction from the current temporary allowance.
Anthropic later acknowledged this comparison after users highlighted it publicly.
Temporary 50% Boost Ends September 13
The current higher allowance has been available for months.
Anthropic originally introduced significant Claude Code usage increases earlier in 2026 and subsequently extended temporary weekly limits multiple times as demand remained high.
The latest announcement establishes September 14 as the point at which the temporary 50% increase will end.
Key Claude Code Dates
| Date | Development |
|---|---|
| May 6, 2026 | Anthropic doubled Claude Code five-hour rate limits |
| May-August 2026 | Weekly usage boosts continued / were extended |
| August 29, 2026 | Permanent +25% weekly increase announced |
| September 13, 2026 | Last day of current temporary boost |
| September 14, 2026 | New permanent limits begin |
Anthropic’s May announcement said the company was increasing Claude Code capacity after securing additional computing resources, including a major partnership with SpaceX.
Anthropic Previously Doubled Claude Code Rate Limits
The latest weekly-limit change should not be confused with another major usage increase Anthropic announced in May.
At that time, the company doubled Claude Code’s five-hour rate limits for Pro, Max, Team and seat-based Enterprise plans.
Anthropic also removed peak-hour reductions for Pro and Max customers.
Anthropic’s 2026 Claude Code Changes
| Change | Details |
|---|---|
| Five-hour rate limits | Doubled in May |
| Peak-hour reduction | Removed for Pro and Max |
| Weekly temporary boost | 50% above baseline |
| New permanent weekly boost | 25% above baseline |
| Current temporary boost ends | September 13 |
| New permanent limits | September 14 |
The different limits operate over different time periods, so users should not assume that the 25% weekly change represents a 25% change in every Claude Code usage restriction.
Usage Is Not Measured Simply By Number Of Prompts
Claude Code’s allowance does not translate into a fixed number of prompts.
Anthropic says usage depends on factors such as conversation length, model selection, tool usage and the effort level selected by the user.
That means two developers could perform very different amounts of work while consuming the same general usage allowance.
Factors That Affect Claude Code Usage
| Factor | Potential Effect |
|---|---|
| Model selected | More capable models can consume more usage |
| Conversation length | Longer context increases consumption |
| Tool use | Additional tool calls use capacity |
| Effort level | Higher reasoning effort can consume more |
| Task complexity | Complex coding tasks require more computation |
| Number of turns | More interactions increase usage |
This makes it difficult to translate Anthropic’s percentage changes into a precise number of coding hours or tasks.
Pro, Max, Team And Enterprise Users Are Affected
The permanent increase applies across Anthropic’s main paid Claude Code plans.
For Pro and Max users, Claude and Claude Code usage is tied to the same broader account usage system.
Anthropic’s help documentation says usage across Claude product surfaces can count toward the same usage limits, while different subscription tiers have different allowances.
Plans Covered By The Change
| Plan | Permanent Weekly Increase |
|---|---|
| Pro | +25% |
| Max | +25% |
| Team | +25% |
| Seat-based Enterprise | +25% |
| Free | Not included in this announcement |
| API users | Separate metering |
The change is therefore primarily relevant to developers who use Claude Code through Anthropic’s paid consumer, team or seat-based enterprise subscriptions.
Anthropic Says More Usage Changes Are Coming
The company has indicated that the September adjustment is not the end of its changes to Claude usage.
Anthropic said it is working on features intended to give users more visibility and control over their usage.
The company has not yet disclosed the full details of those changes.
This could mean better usage tracking, improved quota information or additional controls over how users consume their allowance.
Potential Usage-Control Improvements
Current experience
│
▼
Usage allowance
│
▼
Limited visibility
│
▼
User reaches limit
│
▼
Work interrupted
Potential future experience
│
▼
Real-time usage visibility
│
├── Remaining allowance
├── Consumption trends
├── Model usage
└── Usage controls
│
▼
Better planning
Anthropic has not confirmed exactly which features will be introduced.
Why Anthropic Is Adjusting The Limits
The company says it has been evaluating how much Claude Code capacity it can sustainably provide.
The adjustment comes during a period of exceptionally strong demand for AI coding tools.
Anthropic has been expanding its computing infrastructure aggressively to support Claude and Claude Code.
In May, the company said its SpaceX partnership would provide access to more than 300 megawatts of additional computing capacity and more than 220,000 NVIDIA GPUs within the month.
More recently, Reuters reported that Anthropic agreed to spend $45 billion to rent AI computing capacity from Nscale over six years, representing about 460 megawatts of power capacity.
Anthropic’s Compute Expansion
| Infrastructure Development | Reported Scale |
|---|---|
| SpaceX capacity announced in May | >300 MW |
| GPUs cited in May | >220,000 |
| Nscale agreement | $45 billion |
| Nscale power capacity | ~460 MW |
| Nscale agreement duration | 6 years |
| Additional infrastructure deals | Amazon, Google, Microsoft/NVIDIA and others |
The infrastructure expansion demonstrates how usage limits are closely connected to the availability and cost of AI computing capacity.
Claude Code Demand Is Rising
Claude Code has become an increasingly important part of Anthropic’s product strategy.
The company has been positioning Claude not simply as a conversational AI assistant but as an agent capable of performing complex software-development tasks.
Developers can use Claude Code to inspect repositories, modify files, execute commands, debug problems and work through multi-step coding projects.
That deeper level of interaction can consume substantially more computing resources than conventional chatbot conversations.
Traditional Chat Vs Claude Code
| Traditional AI Chat | Claude Code |
|---|---|
| Text responses | Software-development actions |
| Short interactions | Long-running sessions |
| Limited tool use | Extensive tool use |
| Smaller context requirements | Large project context |
| Primarily informational | Agentic execution |
| Lower compute intensity | Potentially higher compute intensity |
The growing popularity of agentic coding workflows helps explain why usage capacity has become a major issue for Anthropic.
Developers Are Already Sensitive To Usage Limits
Claude Code users have frequently discussed usage limits across developer communities.
When an AI coding assistant becomes part of a developer’s daily workflow, hitting a quota can interrupt active projects.
This makes changes to weekly allowances more consequential than ordinary feature updates.
A developer working on a large codebase could potentially consume a significant portion of the allowance during a particularly intensive debugging or refactoring session.
What A Weekly Limit Means For Developers
Weekly quota
│
├── Feature development
├── Debugging
├── Refactoring
├── Testing
├── Code review
└── Documentation
│
▼
Total usage
│
▼
Weekly limit
Users therefore need to manage model choice, effort and session length to make the available capacity last longer.
Opus Can Consume More Usage
Anthropic’s Claude Code documentation says Sonnet is generally the default choice for most coding work, while Opus is intended for harder reasoning tasks such as major refactoring, difficult debugging and architectural decisions.
The documentation also notes that Opus uses meaningfully more quota than Sonnet.
This means model selection could become increasingly important after the September adjustment.
Model Selection Strategy
| Model Type | Typical Use |
|---|---|
| Haiku | Quick lookups and mechanical tasks |
| Sonnet | Most coding tasks |
| Opus | Complex reasoning and architecture |
| Higher effort | More intensive reasoning |
| Lower effort | Routine tasks |
Developers who use Opus extensively could feel the effective quota change more strongly than users who primarily rely on lighter models.
Anthropic Is Competing For AI Coding Users
The Claude Code limit change comes amid intensifying competition among AI coding products.
OpenAI’s Codex, Google’s coding agents and other AI development tools are competing for professional developers.
OpenAI has also been testing more persistent behavior in Codex, while Anthropic continues to expand Claude Code’s agentic capabilities.
The amount of compute available to each service is becoming an important competitive factor.
AI Coding Competition
| Platform | AI Coding Offering |
|---|---|
| Anthropic | Claude Code |
| OpenAI | Codex |
| Gemini coding tools | |
| Microsoft | Copilot |
| Other developers | Agentic coding platforms |
Developers may increasingly compare not only model quality but also usage limits, pricing, speed and reliability.
The Change Is Better Than Returning To The Original Baseline
Although current users face a reduction from today’s temporary allowance, the new permanent limit is still higher than the original standard.
A user who previously received 100 units will receive 125 units after September 14.
So the long-term position is technically an improvement over the pre-promotion baseline.
The controversy stems from the fact that Anthropic allowed users to become accustomed to 150 units before reducing that amount to 125.
Three-Stage Comparison
Original
100
████████████████████
Current temporary
150
██████████████████████████████
Permanent from Sept. 14
125
█████████████████████████
This explains why Anthropic can accurately describe the change as a permanent 25% increase while current users can accurately describe it as a roughly 17% reduction.
What It Means For Claude Code Users
For developers currently using the temporary 50% boost, September 14 will mean less weekly capacity than they have today.
Users who were already close to their limits may need to adjust their workflows.
That could include using Sonnet instead of Opus for routine tasks, lowering effort settings when appropriate, starting fresh sessions between unrelated projects and using context-management tools more efficiently.
Anthropic’s documentation recommends /clear for starting new tasks and /compact when continuing long sessions, both of which can help manage context and usage.
The Bigger Picture
Anthropic’s decision to permanently raise Claude Code’s standard weekly limits by 25% reflects both the rapid growth of AI coding and the enormous computing requirements behind agentic software development. The catch is that users currently receiving a temporary 50% boost will see their available capacity fall by about 17% when the promotion ends on September 13. The new 125% baseline will nevertheless remain 25% above the original standard allowance.
The change also highlights the growing connection between AI product limits and infrastructure availability. Anthropic has been securing massive amounts of computing capacity, including a SpaceX partnership and a reported $45 billion Nscale agreement, as demand for Claude and Claude Code grows. The company now faces the challenge of balancing developer demand, infrastructure costs and a sustainable usage model while competing with other AI coding platforms.
Looking Ahead
The key date for Claude Code users will be September 14, when the temporary 50% boost ends and the permanent 25% increase over the original baseline takes effect. Developers who have become accustomed to the current higher allowance should expect a reduction in weekly capacity and may need to adjust how they allocate usage across models, projects and long coding sessions.
Anthropic’s promise of better visibility and control over usage could become just as important as the quota itself. If the company can provide clearer information about consumption and help users manage their remaining capacity, the lower limit relative to today’s promotion may be easier to accommodate. Meanwhile, Anthropic’s continued investment in compute suggests that usage limits will remain a moving target as Claude Code demand and AI-agent workloads continue to grow.
Get the day’s top stories in your inbox
One concise email. No spam, unsubscribe anytime.
