Anthropic reduced interaction limits with the Claude bot during peak hours

What has changed in Claude’s operation?

Anthropic has revised the rules for using time limits for users of the Claude chatbot. This decision aims to distribute computing power more evenly among users during peak loads.

What changed? How it looks
Five‑hour session limit (Free, Pro, Max)
During peak hours it can be exhausted faster. Outside those times it is used as usual.
Weekly limits remain unchanged – the total quota per week does not change.

When does this happen?
- Peak interval: 05:00–11:00 Pacific Time (16:00–22:00 Moscow).

- During that period the “five hours” may run out faster than on a normal day.

According to technical team member Tarik Shihipar, about 7 % of users – mainly those on paid plans – will encounter the new limits.

Why is this important?
Anthropic links operating hours with token consumption but does not disclose the exact “tokens‑hours” ratio. Therefore it’s hard for users to plan in advance how many tokens they’ll spend. The official documentation states that influencing factors include:

- Length and complexity of dialogues;
- Features used;
- Selected Claude model;
- Subscription tier (Free, Pro, Max, Team).

How to react?
- Run high‑token‑consumption background tasks outside peak hours.
- During low demand periods available capacity is higher, so total weekly usage stays the same – only its distribution changes.

What do users see?
Claude’s interface has a monitoring panel showing:

- Current consumption of the five‑hour limits;
- Remaining weekly volume.

When a limit is exhausted, access to the service is blocked until the user pays for additional use.

Access to Anthropic services
1. Developer API – pay per token: each processed and generated text incurs a fixed price.

2. Claude subscription – users pay fixed rates, not per operation. The internal order of applying limits remains closed.

Pricing
Plan | Cost
Free | Free
Pro | $20/month
Max 5x | $100/month
Max 20x | $200/month

Thus Anthropic optimizes the use of computing resources while maintaining flexibility and transparency for end users.