Sending a "Hello" to Claude on the standard Pro plan isn't just small talk - it costs roughly 4% of your session limit, according to users tracking their usage.
The math is uncomfortable: at 4% per message, a handful of warmup messages before getting to your actual task can burn 20% of a 5-hour session before you've done anything meaningful. Claude's usage limits work by counting tokens - the basic unit AI models use to process text, where roughly 750 words equals about 1,000 tokens. What most users aren't accounting for is that every message, no matter how short, also triggers a large hidden system prompt that loads behind the scenes. That overhead is where most of the 4% goes.
This isn't a new behavior - it's how Claude has always worked. But it's becoming more visible as users hit limits during intensive sessions and trace backward to figure out where their allocation went. The short version: a casual opener costs proportionally more than it should because session cost is driven by background processing, not the length of your message.
The practical fix is simple: start with your actual request, not a greeting. Claude doesn't need a warm-up. Anthropic hasn't announced any changes to how session limits are metered, so this is a usage pattern issue, not a bug that's going to get patched.