Link to study: https://www.biorxiv.org/content/10.1101/2021.04.05.437224v1.full.pdf
One year later, 70% of people in Wuhan, China who had covid still have detectable levels of antibodies. However, their levels of antibodies have decreased by about two-thirds, leaving one-third of the levels they had at 1-2 months. The levels they have now probably offer some protection, but not a lot, especially against the new variants.
In this study, those who had higher antibody levels to start out with were more likely to still have detectable antibodies a year later, which bodes well for the RNA vaccines, which produce higher levels than does natural exposure. But most likely, vaccinated people will need to take a booster in order to maintain protection against symptomatic infection, although I would expect protection against severe symptoms to last longer than protection against mild symptoms.
Of course, the people in this study probably still have memory B-cells that could make more antibodies within a few days if they were exposed to the virus again. They probably also have memory T-cells. So on average, I would expect them to have lower rates of severe covid than people who had no prior exposure. But we have seen cases where some individuals with poor initial immune responses end up with more severe symptoms the second time than the first time, so it's not a guarantee.