Pro Max 5x quota exhausted in 1.5 hours despite moderate usage (github.com)

Article summary

A user reported that their Pro Max 5x quota was exhausted in 1.5 hours despite moderate usage, which was unexpected given their previous experience with the service. The user investigated and found that cache_read tokens may be counting at full rate against the rate limit, negating the cost benefit of prompt caching. This issue has significant implications for users who rely on the service for development work. The user provided detailed analysis and suggested improvements to address the problem.

Main themes

• Quota exhaustion
• Cache read token accounting
• Rate limiting
• Prompt caching
• Subscription models
• AI-powered development tools

What commenters say

• The quota exhaustion issue is a major problem that needs to be addressed by the company to maintain user trust and satisfaction.
• Some users are considering switching to alternative services due to the quota issues and perceived lack of transparency from the company.
• Optimizing setup and workflow can significantly reduce token usage and mitigate quota exhaustion, but this may not be feasible for all users.
• The company's response to the issue has been inadequate, and users are calling for more transparency and action to resolve the problem.
• Some users have found alternative services to be more generous with their quotas and more effective for their needs, leading to a loss of customers for the original company.
• The issue may be related to the company's attempts to scale and manage their services, but this does not excuse the impact on users.
• Users are divided on the effectiveness of different AI-powered development tools, with some preferring one service over others for specific tasks and workflows.
• The lack of clarity around what constitutes a 'request' and how quotas are calculated is a major source of frustration for users and needs to be addressed.