Anthropic doubles Claude usage outside peak hours — but it won’t last forever

AI labs keep searching for ways to pull developers deeper into their ecosystems. The latest move comes from Anthropic, which says it will double the usage limits for its Claude assistant during off-peak hours — a short-term perk that perhaps says more about competition for developer attention than about generosity.

Anthropic said the promotion runs for two weeks and applies to all company plans, including the free, Pro, Max, and Team tiers. Usage limits are doubled throughout the weekend and outside weekday peak hours, which means the higher allowance applies before 8 a.m. Eastern and after 2 p.m. Eastern on weekdays (5-11 a.m. Pacific), and all day on Saturdays and Sundays. The bonus capacity is applied automatically.

The increase applies to Claude’s rolling five-hour usage window — the system Anthropic uses to cap the number of prompts users can run before temporarily reaching a limit. This means that developers working with Claude — including through tools such as Claude Code — can now run more prompts and longer sessions during those windows without hitting their normal limits.

However, the extra capacity applies only to Anthropic’s own Claude surfaces, including the Claude app (web, desktop, and mobile), Cowork, Claude Code, Claude for Excel, and Claude for PowerPoint. The promotion appears not to extend to the Claude API — a clear sign that the goal is to encourage developers to spend more time inside Anthropic’s own products, rather than external applications built on the model.

While the company describes the change as a “thank you” to users, the timing lands amid an increasingly aggressive race among AI model providers to attract developers and anchor new software ecosystems around their tools.

A burst of extra capacity

The promotion hints at how AI companies are managing the growing demand placed on their infrastructure. Because the change applies automatically, users don’t need to toggle any settings or upgrade a plan. The additional usage is simply available during those time windows.

But allowing more prompts during quieter hours may help smooth demand across Anthropic’s infrastructure. AI models require significant computing resources, and spikes in usage during working hours can place a heavy load on those systems.

Encouraging developers to run jobs outside those windows spreads traffic more evenly, a load-balancing incentive with precedent in cloud computing and data-center operations.

There is also a capacity calculation behind the move. While shifting activity away from peak hours can ease pressure on infrastructure during the busiest parts of the day, the same dynamic works in the opposite direction: the large GPU clusters that run modern AI systems operate continuously, whether they are fully utilized or not. Incentives that encourage heavier use during quieter periods can therefore help fill otherwise idle compute, extracting more work from existing hardware.

Still, the promotion also serves a second key purpose: giving developers more opportunities to build habits around Claude. Once a project becomes tightly linked to a particular model — through prompts, integrations, or fine-tuned systems — switching later can require significant engineering work.

So while Anthropic positions this move as a simple “thank-you,” it may also function as a carrot on a stick — and, to some developers, it might even feel like a bait-and-switch. For two weeks, users get used to running more prompts and working with looser limits. When the promotion ends and the normal allowance returns, that contrast could make the standard cap feel tighter — nudging heavier usage or encouraging upgrades to higher-paid tiers.

The battle for developers

Anthropic’s move arrives during a broader push by major AI labs to recruit developers, particularly those building open-source tools.

Recent initiatives from both Anthropic and OpenAI have offered free or discounted access to their models to maintainers of widely used open-source projects. The idea is simple: if influential tools and libraries adopt a particular model early, entire developer communities may follow.

Anthropic has also been adjusting Claude’s pricing in ways that make heavier use easier. The company recently removed the surcharge on Claude Opus 4.6 and Sonnet 4.6 for requests above 200,000 tokens, meaning pricing now stays flat across the full one-million-token context window.

However, the economics of large language models make long-term “generosity” unlikely. Running frontier AI systems requires vast amounts of computing power, specialized hardware, and energy.

For now, Anthropic’s offer gives users more room to work with Claude during quieter hours. The promotion may be temporary, but the message is clear: AI labs are competing hard to get developers to spend more time — and do more work — inside their ecosystems.