You'll noticeTier: how much it should matter to you
3Useful: my rating, 1 to 5
2Signal: worth watching, 1 to 5
ModelsArea: what it touches
ImprovementsKind: in v2.1.238,
ImprovementsSection of the release
What
The confirmation shown when switching model or reasoning effort now depends on whether prompt caching is on for the current model and the last main-thread request is still within the cache lifetime, rather than a stored flag. The warning about discarding a warm cache appears only when there is one.
Details
Setting DISABLE_PROMPT_CACHING (or the per-model equivalents) suppresses the warning entirely, since caching is off.
Evidence
repl_main_thread
Strings lifted out of the shipped bundle, so the claim above can be checked against them.
