Anthropic's
Hidden Token Cap

The implied cap — tokens burned ÷ utilization — exposes what Anthropic's API actually allows per window. Step-changes reveal when allowances are silently adjusted.

Updated hourly · vmfarms.com

Current Implied Cap

5-Hour Window

329.32M

69.16M burned 21.0% utilized

7-Day (All Models)

2836.84M

1503.53M burned 53.0% utilized

7-Day (Sonnet)

1952.63M

1503.53M burned 77.0% utilized

What is the implied cap? Anthropic publishes utilization percentages but not the underlying token limits. By dividing tokens burned by the utilization fraction, we back-calculate the effective quota cap for each rate-limit window. A sudden jump in the chart below = Anthropic changed your limit.

Implied Token Cap — 90 Days

Step-changes indicate Anthropic silently adjusted quota allowances.

Utilization

5-Hour Window 21.0%

69.16M burned / ~329.32M cap as of 2026-03-27 03:55 UTC

7-Day (All Models) 53.0%

1503.53M burned / ~2836.84M cap as of 2026-03-27 03:55 UTC

7-Day (Sonnet) 77.0%

1503.53M burned / ~1952.63M cap as of 2026-03-27 03:55 UTC

Utilization % — 90 Days

How much of the quota is being consumed per window.

Raw Snapshot

Window	Utilized	Burned	Implied Cap	As Of
5-Hour Window	21.0%	69.16M	329.32M	2026-03-27 03:55 UTC
7-Day (All Models)	53.0%	1503.53M	2836.84M	2026-03-27 03:55 UTC
7-Day (Sonnet)	77.0%	1503.53M	1952.63M	2026-03-27 03:55 UTC

Need fully managed cloud hosting without the markup? We handle the infrastructure — bare metal performance, managed for you.

vmfarms.com →