[BUG] 5-hour session quota is misleading — actual limit is token-based, not time-based
Preflight Checklist
- [x] I have searched existing issues and this hasn't been reported yet
- [x] This is a single bug report (please file separate reports for different bugs)
- [x] I am using the latest version of Claude Code
What's Wrong?
The session is labeled as a 5-hour quota, but in practice the limit is hit based on token/context usage, not elapsed time. On a project with a moderate-sized codebase, system prompts, and memory files, the limit can be reached in under an hour with only a handful of prompts.
The "5hr" label sets a false expectation. Users should be clearly informed that the quota is token-based, and the UI should reflect remaining context, not time.
What Should Happen?
Expected: Session lasts up to 5 hours as advertised
Actual: Session context exhausted in under 1 hour
Error Messages/Logs
Steps to Reproduce
Steps to reproduce:
Open Claude Code on a project with CLAUDE.md, memory files, and a multi-package monorepo
Run 3–5 prompts that involve file reads
Observe context percentage climbing rapidly — well under 1 hour
Claude Model
Sonnet (default)
Is this a regression?
Yes, this worked in a previous version
Last Working Version
_No response_
Claude Code Version
N/A using plugin in VSCode
Platform
Anthropic API
Operating System
macOS
Terminal/Shell
Terminal.app (macOS)
Additional Information
_No response_
This issue has 3 comments on GitHub. Read the full discussion on GitHub ↗