Claude Code v2.1.168

Claude Code now enables extended thinking for models in the "mythos" family (such as claude-mythos-preview) where it was previously disabled. Two distinct thinking budget paths are updated:

• The always-disabled thinking path ([false, 0]) now returns [undefined, 2048] for mythos-family models, activating thinking with a 2048-token budget.
• The main thinking configuration path now checks for mythos-family models first and applies a dedicated 2048-token budget constant (Cx4) before falling through to the general thinking settings.

This means users who have claude-mythos-preview configured as their model will now benefit from extended reasoning where Claude thinks through problems more carefully before responding.

Two new model family identifiers, "fruitcake" and "macaroon", are pre-registered alongside "mythos" in the model-family detection array. These codenames are used by the same thinking-budget routing logic described above, meaning that when models with these internal names ship they will automatically inherit the same 2048-token thinking configuration without requiring a code change.

No user-facing claude-fruitcake- or claude-macaroon- models are available yet; this is infrastructure preparation.

A new helper function was added that finds which advisor model family a given model name belongs to by doing a case-insensitive substring search against the list of valid advisor model families (["opus", "sonnet"]). This replaces ad-hoc string comparisons and makes error messages for invalid advisor choices more consistent.