ChatGPT is not honoring the selected model for scheduled tasks. It always runs the task in GPT-6 Instant unless I retry on my phone with “Use Thinking” enabled; then it uses GPT-6 Thinking. Either way, it never uses the model I selected.
Task: Morning Work Briefing
Confirmed. This is especially bad when 5.6 is the requested model and it forces your thread back to 6.
Hey @ch13fkeefe, thanks for flagging this! Is Morning Work Briefing a personal or team task, and does it run in Chat or Work? Could you share screenshots showing the model and thinking level selected in the task settings and where you’re seeing “GPT-6 Instant” on the scheduled result? Please hide any personal details and add the time and time zone of the November 3 run so I can look into it.
@Shadow1337, thanks for sharing your experience with the GPT-5.6 selection too.
Craig