Manatees (6) and sloths (5-10) are the exceptions.
It's very interesting that it is only cervical vertebrae that remain nearly universally consistent. Lumbar and thoracic numbers vary quite wildly.
I found this really interesting paper hypothesizing that a combo of low metabolism and low activity can lead to a loosening of evolutionary constraint on that number, based on the sloths and manatees being the outliers. Their study also shows bushbabies being the only primate where around 30% of the 80ish specimens studied had a thoracic vertebra that had not fully developed a rib, making it much more like a cervical vertebra. They were said to be the least active and metabolically demanding of the primates, further supporting their hypothesis.