[BUG] 2.1.63 silently defaults to 1M context model without UI indication, causing excessive usage consumption and a BILLING issue.

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?

Version 2.1.63 silently defaults to the 1M context window model without displaying the "(1M context)" label in the UI.

Statusline shows "Opus 4.6", /model checks show "Opus 4.6", asking CC directly says "claude-opus-4-6", but /context reveals claude-opus-4-6 · 41k/1000k tokens — confirming 1M context is active.

Users are unknowingly billed at extended context rates, causing ~5-7x faster usage consumption. This is a 2.1.63 regression — 2.1.62 correctly defaults to 200K.

What Should Happen?

1. Default model should use 200K context window
2. If 1M context is active, UI should display "(1M context)" and /model (and other checks) should clearly reveal 1m context
3. 1M context should be opt-IN, not silent default (unless extra billing free)

Error Messages/Logs

No error — silent misconfiguration.

Evidence from /context command on 2.1.63:
  claude-opus-4-6 · 41k/1000k tokens (4%)

Statusline shows: "Opus 4.6 + ultrathink" (no 1M indicator)

After setting CLAUDE_CODE_DISABLE_1M_CONTEXT=1:
  Context changed from 5% to 28% (same tokens, correct 200K window)

Steps to Reproduce

1. Install Claude Code 2.1.63
2. Start a new session with Opus 4.6
3. Run /model commands, or ask CC its model, observe "Opus 4.6" or "claude-opus-4-6" (no 1M indicator)
4. Run /context
5. Observe output shows claude-opus-4-6 · xk/1000k tokens — 1M context active without indication

To verify:

1. Add to ~/.claude/settings.json:

{ "env": { "CLAUDE_CODE_DISABLE_1M_CONTEXT": "1" } }

1. Start new session
2. Context % jumps from ~5% to ~25% (same tokens, now against 200K window)

Comparison: 2.1.62 and previous defaults to 200K correctly and shows "(1M context)" label when 1M is manually selected.

Claude Model

Opus

Is this a regression?

Yes, this worked in a previous version

Last Working Version

2.1.62

Claude Code Version

2.1.63

Platform

Anthropic API

Operating System

Ubuntu/Debian Linux

Terminal/Shell

Other

Additional Information

<img width="1435" height="934" alt="Image" src="https://github.com/user-attachments/assets/bae5c1ef-5773-4376-820b-b7406fe80977" />