A risk based approach should be employed. For autoclaves and depyrogenation ovens, a tighter requirement is needed to ensure sterility and lethality calculations are correct, and the sensitivity of the calibration can impact the Lo factor calibrated. I would stick to 0.1 or tighter for these applications.
For freezers and fridges, the sensitivity of the product likely doesn’t hinge on tenths of a degree, and a more loose requirement can be defended. If you have a calibration “bath” that goes down to -80 C, then you should still be able to get 0.1 precision - if you rely on manual calibration, then you’d likely have to adjust it accordingly.