[Bug Report] Weekly token usage limit draining abnormally fast immediately after Claude Code Sonnet 4.6 update — using Opus 4.6 on Max x5 plan ($200/month)

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?

Issue 1: Weekly token usage is definitely draining faster than normal

I am currently on the Claude Max x5 plan ($200/month) using Claude Opus 4.6 via Claude Code (Windows, Git Bash terminal). I have confirmed that my weekly token usage is draining significantly faster than it used to. This is not speculation — it is a clearly and definitively observed symptom happening right now. I am requesting Anthropic to investigate and officially confirm the cause.

Additional compounding factor: I am a Korean-language user, which already results in 2x–3x higher token consumption due to tokenization inefficiency (see related issue #26401). This makes the problem even more severe.

---

Issue 2: Max plan pricing multiplier does not match actual usage multiplier

According to Claude's official pricing page, the plan tiers are:

• Max x5 = $100/month
• Max x20 = $200/month

This means upgrading from Max x5 to Max x20 doubles the price ($100 → $200), but should deliver 4x the usage (x5 → x20 = 4x increase in the multiplier).

However, in practice, the actual usable token amount and the real-world experience of upgrading from $100 to $200 only feels like approximately 2x–2.5x improvement — not the 4x that the plan multipliers suggest.

This gap is a serious problem:

• The token limit in practice does not scale proportionally with the price increase
• The subjective experience of upgrading is far less than what the advertised multiplier (x5 → x20) implies
• Paying double the price should yield 4x the usage, but users are only getting roughly 2x–2.5x

I am requesting Anthropic to clearly confirm the actual token allocation difference between the Max x5 ($100) and Max x20 ($200) plans, and explain why the practical difference feels so much smaller than the advertised 4x multiplier.

What Should Happen?

1. Anthropic should officially investigate and confirm why weekly token usage is draining faster than normal for Opus 4.6 users.

1. Anthropic should clearly explain the actual token difference between Max x5 ($100/month) and Max x20 ($200/month), and why upgrading from $100 to $200 only delivers approximately 2x–2.5x improvement instead of the expected 4x (based on x5 → x20 multiplier ratio).

Error Messages/Logs

No explicit error messages — the issue manifests as dramatically accelerated weekly usage limit depletion visible in the Claude Code usage indicator.

Steps to Reproduce

1. Subscribe to Claude Max x5 plan ($200/month)
2. Use Claude Code with Claude Opus 4.6 model via Windows Git Bash terminal
3. Observe that weekly token usage is depleting faster than normal (confirmed, not estimated)
4. Compare against Max x5 ($100/month) experience — the actual usable token amount difference is only approximately 2x–2.5x, not the 4x the plan multiplier ratio (x5 → x20) would suggest

Additional Information

• Plan: Claude Max x5 ($200/month)
• Model in use: Claude Opus 4.6
• OS: Windows
• Terminal: Git Bash
• Language: Korean (primary interaction language)
• Observed weekly token drain: Definitely faster than normal — please confirm
• Observed plan upgrade value: Upgrading from Max x5 ($100) to Max x20 ($200) yields only ~2x–2.5x improvement in practice, not the expected 4x based on official plan multipliers
• Related issue: #26401 — Korean/CJK users already face structural token disadvantage due to tokenization inefficiency

Please explicitly confirm:

1. Why is weekly token consumption faster than normal for Opus 4.6 users?
2. What is the actual token allocation difference between Max x5 ($100) and Max x20 ($200), and why does it not reflect the 4x multiplier ratio?