Hey, as far as I’ve learned, switching to different models in the same conversation results in rereading the context, which wastes tokens. But this doesn’t happen when switching from Agent mode to Ask mode, for example, right? @condor
hi @TRTRTRTR as long as the model remains the same there is no difference in the mode from AI perspective.
A model change e.g. from Anthropic’s Sonnet 4 to OpenAI’s GPT-5 requires us to send the whole context of current chat as new input to the new model. As there was no cache with the new provider for that chat thread it will be considered by the provider as new input and charged accordingly.
1 Like