Skip to content
RCreddit.com·

Claude Pro Token Usage Has Increased Dramatically After the Latest Update. Anyone Else?

AI summary

A user of Claude Pro, specifically the Opus model, has observed a dramatic increase in token usage following a recent update. Previously, they could engage in 1-2 hours of continuous, intensive work without significant issues. Now, they are questioning if others are experiencing a similar rise in token consumption over the past few days, suggesting a change in how the service operates after the latest update.

Time & source
Published
Sep 7, 2026, 04:33
Source type
Dev community
Tier
Community
Source status
Healthy
Tier is a per-source editorial setting, not a per-item score.

Times shown in UTC

More details
First seenSep 7, 2026, 12:00Time zoneUTC · UTC+0
Article

Has anyone else experienced an exponential increase in token usage on Claude Pro after the latest update?

I always try to follow good practices to get the most out of Claude’s usage limits. For example, I avoid leaving unnecessary chats open, and when I’m working on an application and the conversation starts getting too long, I ask Claude to generate a .md file with the necessary context so I can continue the work in a new chat and keep the conversation cleaner.

I’ve been paying for Claude Pro for around 4 months, and throughout that time I’ve almost always worked with Opus. Normally, I could get 1–2 hours of continuous, intensive work without any major issues.

However, lately I’ve noticed that my token usage has increased dramatically. Here’s an example for reference: I gave Claude a task where I only asked it to make two very minor changes, and it consumed more than 40% of my limit, whereas previously a similar task would consume at most around 5%.

What’s even stranger is that I haven’t changed the way I work. I’m still following the same best practices to optimize token usage, and I don’t use any third-party plugins or extensions; I use Claude completely natively through Claude.ai.

I also make sure not to push the context windows too far, since I want to make the most efficient use of my available tokens.

Is anyone else experiencing something similar? Have you noticed a significant increase in token consumption over the past few days?

Source·reddit.com·Full text via RSS