Hacker News

portkey

TechCrunch

<h1>Anthropic Introduces Weekly Rate Limits for Claude Code Subscribers</h1>
<p>Anthropic has recently announced the introduction of weekly rate limits for its Claude Code service, effective August 28, 2025. These changes will impact both Pro and Max subscribers. This decision is part of a broader trend among AI companies to manage usage and computational resource constraints, although it has sparked discussions about fairness and transparency.</p>
<h2>Background:</h2>
<p>AI companies, including Anthropic, have historically offered subscriptions claiming “unlimited” access, only to later impose usage restrictions. This practice can be particularly frustrating for power users who have integrated these tools into their workflows, as it challenges the trust and commitment necessary for full organizational adoption.</p>
<h2>New Limits Overview:</h2>
<p>Under the new structure, Claude Code operates on a 5-hour rolling session model. Weekly quotas impact less than 5% of users who are known for intensive usage patterns, including those running the AI coding tool continuously and sharing accounts. Most users will receive 140-280 hours of Sonnet 4 and 15-35 hours of Opus 4 per week. Max subscribers can multiply their usage capacity with tiers priced at $100 and $200 per month, offering enhanced access including Opus 4.</p>
<h2>Pricing Details:</h2>
<p>Pro plans are priced at $20/month with capabilities for 10-40 prompts per session. API usage charges are standardized across platforms like Anthropic, AWS Bedrock, and Google Vertex AI at $3/$15 per million tokens for Sonnet 4 and $15/$75 for Opus 4. Enterprise users have additional options to streamline control and observability through Portkey’s AI Gateway.</p>
<h2>Criticism and Alternatives:</h2>
<p>The adjustments have been met with criticism from developers, who argue that legitimate power users are unfairly penalized due to the actions of others abusing the system. Yet, Anthropic maintains that rate limits are necessary to ensure system reliability and to prevent policy violations that can disrupt service for all users. Comparatively, similar pricing strategies and restrictions have been adopted by other AI tool providers, including Cursor and Replit.</p>
<p>This change emphasizes the importance of transparent communication and clear expectations regarding usage policies for AI services. As computational demand grows, companies must balance access and performance while maintaining user trust.</p>


Collections

Anthropic has implemented weekly rate limits for Claude Code subscribers starting August 28, 2025, affecting both Pro and Max tiers. The new structure operates on 5-hour rolling sessions with weekly quotas impacting less than 5% of intensive users. Most users receive 140-280 hours of Sonnet 4 and 15-35 hours of Opus 4 weekly, while Max subscribers can access higher tiers at $100-200/month. The changes have drawn criticism from developers who feel legitimate power users are penalized, though Anthropic cites system reliability and policy violation prevention as reasons for the restrictions.

Anthropic Introduces Weekly Rate Limits for Claude Code Subscribers

<p>This post tries to spin a usage cap as “minor” when in reality it signals a deeper issue which is the collapse of the unlimited inference promise. Power users who push innovation forward are now collateral damage. The core problem isn’t resource management, it’s the mismatch between pricing and compute economics. You can’t sell a metered commodity like inference under a flat-rate subscription and expect sustainability. Worse, the opacity around quotas, model routing, and actual usage undermines trust.</p>


<p>This is inevitable. Since computing is a function of electricity which is not a fixed monthly cost, it’s impossible to do subscription unless it’s absurdly expensive such that the average subscriber is not using it enough to get their money’s worth.</p>


<p>And there comes the infra cost wave</p>


<p>As a Claude Code subscriber, I’ve really appreciated the tool’s capabilities and had integrated it deeply into my development workflow. That’s why the announcement about weekly rate limits is pretty frustrating.</p>
<p>I understand the need to manage system load and prevent abuse, but it feels like legitimate, high-volume users are being penalized for the actions of a small minority. Many of us signed up under the impression that the service was “unlimited”—and for those who rely on it professionally, sudden usage caps disrupt more than just convenience; they impact project timelines and team collaboration.</p>
<p>The additional pricing tiers might work for some, but it also feels like paywalls are slowly creeping in where there was once more open access. I really hope Anthropic takes user feedback seriously and considers more flexible or transparent options—maybe usage dashboards, warnings before limits hit, or custom plans for power users.</p>
<p>I still think Claude is one of the best coding AIs out there, but trust and consistency are just as important as performance.</p>
