Anthropic says that its deployment of new weekly rate limits for Claude to limit use among subscribers who perform its AI coding tool, Claude Code, “continuously in the background, 24/7”. Anthropic claims that the rate limits also aim to stop a handful of users who violate Claude’s use policy by sharing accounts and by reselling access to the Claude code.
The new rate limits will come into force on August 28 for subscribers to the pro plan of $ 20 per month of anthropic, as well as its maximum plans of $ 100 and $ 200 per month, the company announced on Monday in subscribers and a position on X.
Anthropic says that its existing limits of use, which reset every five hours, will remain in place. The company also introduces two new weekly rate limits that are reset every seven days; One is an overall use limit, while the other is specific to the most advanced IA model of anthropic, Claude Opus 4. Anthropic says that max subscribers can buy additional use, beyond the rate limit provides, standard API levels.
The announcement comes only a few weeks after Anthropic was quietly introduced the rate limits for Claude Code. The company said at the time that it was aware of the problems but refused to develop more. Although the AI coding tool of Anthropic was a success with developers, the company seems to have trouble serving it largely. The Anthropic State page shows that Claude Code has experienced a partial or major breakdown at least seven times during the last month – perhaps because some power users seem to execute Claude Code without stopping.
“Claude Code has undergone an unprecedented request since its launch,” said anthropogenic spokesperson Friend Rotherham in a Techcrunch e-mail about the weekly rate limits. Rotherham notes that “most users will not notice a difference” and that this limit will affect less than 5% of subscribers, depending on their current use models.
Anthropic said to Techcrunch that most professional users can expect 40 to 80 hours from Sonnet 4 via Claude Code within their weekly rate limits. Subscribers at the maximum plan of $ 100 per month of anthropic can expect 140 to 280 hours from Sonnet 4 and 15 to 35 hours of opus 4. and the maximum plan subscribers of $ 200 per month of anthropic can expect 240 to 480 hours of Sonnet 4 and 24 to 40 hours of opus 4.
The company notes that use may vary depending on the size of the code base and other factors, however, it is not clear how Anthropic was measuring use here. Anthropic claims that the maximum plan of $ 200 offers 20 times more use than the professional plan – but on the basis of the updated figures, subscribers now only get 6x as many hours of Code Claude as pro users.
Techcrunch event
San Francisco
|
October 27-29, 2025
It is possible that the 20x figure always applies when measured in tokens or calculation, but the company did not immediately clarified.
Anthropic has already declared that it was very limited with regard to IT resources, which seems to be the case for most providers of AI models today. Most AI companies arise to provide new online AI data centers to meet the massive service and training requirements for their AI models.
Several providers of AI coding tools revisit the pricing strategy around their products. In June, the company behind Cursor, Anysphere, changed the way it assessed the use of its professional plan of $ 20 per month to limit energy users to abuse the plan. However, Anysphere apologized later for having badly communicated these changes, which led some users to pay more than what they expected. Another AI coding tool provider, replaces, also made similar price changes in June.
In an e-mail to Claude subscribers, Anthropic says that he is committed to “supporting long-term use cases through other options in the future”. However, the company says that these rate limits will help them maintain reliable short -term services.
Update 4:45 pm: This story has been updated to include comments from an anthropogenic spokesperson and more details on the weekly limits of anthropic.
