Rate limit vs quota
A rate limit caps how fast you can call an AI model, for example requests or tokens per minute. A quota caps how much you may use or reserve in total, such as a monthly spend limit or the capacity assigned to your account in a region.
Why it matters for companies
When agents start failing with HTTP 429 errors, the fix depends on which limit was hit. A larger monthly allowance does nothing for a per-minute ceiling, and the reverse is also true.
Vendors name these limits differently. Anthropic describes two types for the Claude API: spend limits, a maximum monthly cost, and rate limits, measured in requests, input tokens and output tokens per minute for each model class. Microsoft calls Azure PTU quota a policy limit per subscription and region, and notes that quota does not guarantee capacity.
Example
A team's agents fail every morning at 10:00 with 429 errors. The logs show input tokens per minute at the rate limit while the monthly allowance is far from used. Spreading the load across more providers fixes it, a bigger monthly quota would not.
How Standard Compute handles it
Point Claude Code and Claude Desktop at Standard Compute and your team stops hitting 5-hour and weekly caps: agents run until the monthly budget is used. Claude is served across several clouds, so one provider's quota or outage doesn't stop your work.