What is the interpretation of the REML criterion when running lmer?

Question

I wanted to ask how to interpret the displayed REML criterion at convergence when running a HLM model using lmer function in lme4 package? Is it better when it is smaller or larger? Is it comparable among different models for the same outcome variable?

Here is an example of the model:

hlm_example <- lmer(BSMMAT01 ~ 1 + (1 | IDSCHOOL), data=data)
summary(hlm_example)

Model Output