Opus 5.5 becomes the Opus choice in the model picker, defaulting to medium effort

What

Claude Opus 5.5, model id claude-opus-5-5, now fills the Opus slot in the model catalog bundled with Claude Code. The catalog is the built-in list of models the model picker offers, and this build carries catalog version 994. The in-app release notes describe the same change.

• The in-app release notes now start with a ## 2.1.280 section. Its first line says Claude Opus 5.5 (claude-opus-5-5) was added and is now the default Opus model. There is no ## 2.1.281 section in this build.
• The picker's quick-select entry with short name "Opus" and description "For complex tasks" now points to claude-opus-5-5 / "Opus 5.5". Before, it pointed to claude-opus-5 / "Opus 5".
• The static model table shows the display name "Opus 5.5", and there is a Vertex region environment variable named VERTEX_REGION_CLAUDE_5_5_OPUS.
• The per-model thinking defaults (thinking_by_model) now list claude-opus-5-5 on every surface. Thinking effort is how much work the model puts into a reply before answering.
• On the cc surface and in the "chat" preset, claude-opus-5-5 gets effort "medium", while claude-opus-5 keeps "high".
• A bundled model_selector_state sample for cc and ccd shows the same switch to medium, although its top-level model is still claude-opus-5.
• In the catalog's claude-opus-5-5 entry, the recommended "Default" badge in the effort options moved from High to Medium. The same badge move happened for claude-fable-5-1 under the "chat" selector config.
• A separate list that uses effort_and_mode gains a claude-opus-5 entry with effort "high" and mode "auto", placed just before claude-opus-4-8.

Why

Picking Opus in the model picker now selects Opus 5.5, and it starts at medium effort rather than high. The catalog's own effort text says higher effort gives more thorough responses but takes longer and uses your limits faster. If replies feel less thorough than they did with Opus 5, check the effort setting.