[BUG] Max 20x — 50% weekly limit increase (May 13) not reflected in actual usage capacity

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?

Plan: Max 20x ($200/month)
Environment: Windows 11, Claude Code CLI + VS Code, primarily Opus 4.6

On May 13, Anthropic announced a 50% increase in Claude Code weekly limits for all Pro, Max, Team, and Enterprise users (through July 13). I'm on the Max 20x plan. After two full weekly cycles post-announcement, my usage capacity is identical to before — I hit 98% at the same rate, with the same usage patterns.

Using phuryn/claude-usage to track local JSONL logs, here are my weekly totals (API cost equivalent):

| Week | Est. API Cost | Turns | Result |
|------|-------------|-------|--------|
| Apr 20-26 (pre-increase) | $3,098 | 6,778 | Hit limit |
| Apr 27 - May 3 (pre-increase) | $4,129 | 8,144 | Hit limit |
| May 4-10 (pre-increase) | $3,225 | 8,277 | Hit limit |
| May 11-17 (post 50% increase) | $4,411 | 11,193 | Hit limit at 98% |
| May 18-22 (current, incomplete) | $3,689 | 8,923 | On track to hit limit |

Additional data point: My 30-day dashboard shows 8.19B cache_read tokens vs only 906.5K input tokens. Cache reads represent ~99.99% of my input. If cache_read_input_tokens are being counted at full rate instead of the documented 1/10 rate (as suggested in #58557 and #59872), this alone would explain why the 50% increase has zero visible effect.

Possibly related: #58557 and #5

<img width="1725" height="1310" alt="Image" src="https://github.com/user-attachments/assets/50086384-c771-41cc-8f44-4ca963633be7" />
<img width="1698" height="1246" alt="Image" src="https://github.com/user-attachments/assets/1ead34f2-f804-473b-b3b3-1dad7f154921" />

9587

What Should Happen?

My Max 20x weekly limit should reflect the announced 50% increase. Either:

1. The increase was not applied to my account, or
2. A concurrent change (possibly cache_read counting at full rate) increased consumption rate, nullifying the increase

Request: Could someone from the team review my account's actual weekly limit allocation and confirm whether the 50% boost is active? And verify how cache_read_input_tokens are being counted against my limit?

Error Messages/Logs

Weekly usage data from phuryn/claude-usage (local JSONL analysis):

Week Apr 20-26:     $3,098 API equivalent — hit limit
Week Apr 27-May 3:  $4,129 API equivalent — hit limit
Week May 4-10:      $3,225 API equivalent — hit limit
Week May 11-17:     $4,411 API equivalent — hit limit at 98% (POST 50% increase)
Week May 18-22:     $3,689 API equivalent — on track to hit limit (incomplete)

30-day totals: 257 sessions, 42.8K turns, 8.19B cache_read, $6,134 est. API cost

Steps to Reproduce

1. Subscribe to Max 20x plan ($200/month)
2. Use Claude Code normally (CLI + VS Code, primarily Opus 4.6)
3. Track usage weekly with phuryn/claude-usage
4. Compare weekly capacity before May 13 (pre 50% increase) vs after
5. Observe identical ceiling — no increase visible

Claude Model

Opus

Is this a regression?

Yes, this worked in a previous version

Last Working Version

N/A the 50% increase never took effect on my account

Claude Code Version

2.1.142

Platform

Anthropic API

Operating System

Windows

Terminal/Shell

Other

Additional Information

<img width="1725" height="1310" alt="Image" src="https://github.com/user-attachments/assets/099f628a-51cb-40e4-b8ae-c81a7ae63a4e" />
<img width="1698" height="1246" alt="Image" src="https://github.com/user-attachments/assets/bb86d4d3-5e2d-4b8f-ae7d-9089ecc440e8" />