[BUG] Opus 4.6 at 1m context actually is Opus 4.5

Resolved 💬 5 comments Opened Mar 29, 2026 by nanopink Closed Apr 1, 2026

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?

Selecting "Opus 4.6" in Claude Code actually loads claude-opus-4-5-20251101 a completely different, older model.

Users are being served Opus 4.5 while the UI claims it's Opus 4.6. This likely explains the quality degradation and verbose "thinking out loud" behavior users have been reporting, they were never running Opus 4.6 to begin with.

!Image

What Should Happen?

Selecting Opus 4.6 should load the actual Opus 4.6 model (claude-opus-4-6).
Users should receive the model they select.

Steps to Reproduce

  1. Open Claude Code
  2. Run /model or switch from Sonnet 4.6 to Opus 4.6
  3. Observe the model string: claude-opus-4-5-20251101
  4. Expected claude-opus-4-6-* and actually got claude-opus-4-5-20251101

Claude Model

Opus

Is this a regression?

Yes, this worked in a previous version

Last Working Version

Unknown (Before 1m was forced by default and 400k disappeared)

Claude Code Version

2.1.87

Platform

Anthropic API

Operating System

Windows

Terminal/Shell

VS Code integrated terminal

Additional Information

We have been deceived.
There was a sharp model quality decrease when 1m was forced by the update.
And this is because 4.6 at 1m is actually 4.5.

View original on GitHub ↗

This issue has 5 comments on GitHub. Read the full discussion on GitHub ↗