[BUG] Abnormal Token Consumption Rate in Claude Code - Pro Plan

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?

Environment:

• Tool: Claude Code in VS Code
• Plan: Pro (Individual)
• Models Used: Claude Opus 4.8, Claude Sonnet
• Date: Today (June 7, 2026)
• Time Window: 8:00 AM - 9:45 AM (1h 45min)

Issue Description:
Experiencing unusually high token consumption rate that appears disproportionate to actual workload and activity level.

Consumption Data:

• Daily Usage After 1h 45min: 59%
• Weekly Usage: 6%
• Activity: Basic web page development (HTML/CSS/JavaScript)
• No complex multi-file reasoning or intensive operations
• No unusual artifact generation or API calls beyond normal VS Code integration

Expected Behavior:
Token consumption should be proportional to actual code complexity and reasoning required. Light web development should not consume 59% of daily allocation in under 2 hours.

Actual Behavior:
At this consumption rate, daily token allocation would be depleted in approximately 3 hours of normal coding work.

Steps to Reproduce:

1. Open VS Code with Claude Code extension
2. Perform basic web development tasks (writing HTML/CSS/JS)
3. Monitor token usage via /usage command
4. Observe disproportionately high consumption relative to task complexity

Additional Context:

• Verified usage data in Claude Console
• No unauthorized access or suspicious activity detected
• Issue persists across both Opus and Sonnet model selection
• Token counting appears to be the issue, not user behavior

Affected Service:
Recent incident affecting Claude Code and multiple models (resolved at 4:28 UTC today) may be related.

Request:
Investigate whether:

1. Token counting is accurate for Claude Code sessions
2. Recent service incident caused persistent metering issues
3. System prompts or tool calls are being over-counted

What Should Happen?

...

Error Messages/Logs

Steps to Reproduce

...

Claude Model

None

Is this a regression?

Yes, this worked in a previous version

Last Working Version

_No response_

Claude Code Version

Claude 1.11187.4 (584005) 2026-06-05T20:15:04.000Z

Platform

Anthropic API

Operating System

Windows

Terminal/Shell

VS Code integrated terminal

Additional Information

_No response_