# Feature Request: Add Usage Metrics Transparency to Claude Code
## Summary
Claude Code should display real-time usage metrics to help users understand their current usage limits and which model (Opus 4 vs Sonnet 4) is
actively being used.
## Current Behavior
- No visibility into current usage percentage
- No indication when model switches from Opus 4 to Sonnet 4
- Users cannot plan their work based on remaining capacity
- Model switches happen silently at 20% usage threshold
## Desired Behavior
Claude Code should provide:
- Usage Dashboard showing:
- Current usage percentage (e.g., "15% of daily limit used")
- Active model indicator (e.g., "Currently using: Opus 4")
- Remaining capacity before model switch
- Status Indicators:
- Visual indicator in the interface showing active model
- Warning when approaching the 20% threshold
- Notification when model switch occurs
- Usage Commands:
/usagecommand to check current statistics/modelcommand enhancement to show why current model is active
## Benefits
- Better Planning: Users can prioritize complex tasks when Opus 4 is available
- Transparency: Users understand what's happening behind the scenes
- User Control: Ability to make informed decisions about task execution
- Improved Experience: No surprise model switches during critical work
## Proposed Implementation
- Add usage metrics to the session context
- Display metrics in the interface (status bar or dedicated panel)
- Include usage info in the
/modelcommand output - Optional: Allow users to set custom threshold warnings
## Example Interface
````
Claude Code Status:
Model: Opus 4 (15% of 20% limit used)
Daily Usage: 2.5k/15k tokens
Next Model: Sonnet 4 (after 20%)
## User Story
As a Claude Code user, I want to see my current usage metrics so that I can plan my work effectively and understand which model is serving my
requests.
---
This feature would significantly improve the Claude Code user experience by providing transparency about usage limits and model selection.
This issue has 9 comments on GitHub. Read the full discussion on GitHub ↗