Opus-specific rate limit is not reflected in the usage meter — misleading UX

Summary

When using Claude Code on the Pro plan with the Opus model (high effort), users can hit a rate limit error even when the displayed usage meters (weekly and 5-hour rolling window) show plenty of remaining capacity. The Opus-specific sub-quota is not surfaced in the UI at all, which causes significant user confusion.

Steps to reproduce

1. Use a Pro plan subscription
2. Start a Claude Code session with --model claude-opus-4-8 --effort high
3. Continue working until the Opus-specific quota is exhausted
4. Observe: "Usage limit reached" error appears even though:

• Weekly usage meter shows well under the limit
• 5-hour rolling window shows well under the limit

Expected behavior

The usage meter should display Opus-specific quota separately, or at minimum indicate which sub-quota caused the limit. Example:

Overall usage:  ████░░░░░░ 40%
Opus quota:     ██████████ 100% ← limit hit here

Actual behavior

All meters show "plenty of capacity," yet the user receives a rate limit error. There is no indication that a model-specific sub-quota exists or is being tracked.

Impact

• Users are confused and may incorrectly assume there is a billing issue
• Users may purchase additional API credits thinking that will resolve the limit (it does not, since the Pro plan uses time-based rolling quotas, not credit-based limits)
• Users cannot make informed decisions about which model to use if the sub-quota is not visible

Suggestion

• Add a per-model quota indicator to the usage display (even if approximate)
• Or: when a rate limit is hit due to a model-specific sub-quota, clearly state in the error message which quota was exhausted and when it resets (e.g., "Opus quota resets in 2h 15m — switch to Sonnet to continue")

Environment

• Plan: Pro
• Model: claude-opus-4-8
• Effort: high
• Platform: macOS (darwin)
• Claude Code version: 2.1.187