{"version":"2.1.300","anchor":"model-capabilities-add-thinkingalwayson","canonical_anchor":"model-capabilities-add-thinkingalwayson","heading":"Models can now be marked as having thinking that is always on","tier":"internal","area":"Thinking","scope":"individual","heads_up":false,"url":"https:\/\/changelogs.core-directive.com\/v\/2.1.300\/e\/model-capabilities-add-thinkingalwayson","release_url":"https:\/\/changelogs.core-directive.com\/v\/2.1.300","markdown":"### Models can now be marked as having thinking that is always on\n\nClaude Code's record of what a model can do now includes whether its thinking is always on\n\n**Unclear.** Which models are marked as having thinking always on is not stated.\n\n**What**\n\nClaude Code keeps a record of what each model can do. The part about thinking (the model reasoning step by step before it answers) already covered whether thinking is supported, whether the model refuses requests with thinking turned off, and whether it limits effort when thinking is off. It now also has an \"always on\" entry.\n\n**Why**\n\nSome models have thinking that cannot be turned off. Claude Code can now record that about a model.\n\n- Area: Thinking\n- Tier: Under the hood\n- Useful: 2\/5\n- Signal: 4\/5\n- Scope: individual\n- Heads-up: no"}