QUBO variable vs number qubits estimation

I was wondering if there was a way to understand the relationship between the number of logical variables of the QUBO I'm trying to run on D-Wave's advantage system. I know there isn't an exact correlation between QUBO variables and number of qubits, and I know that the minor embedding process (through D-Wave's Minorminor method) accounts for this by using extra qubits to relate variables together, but I was wondering if there was a way to predict the maximum size of a QUBO that is able to run on a quantum annealer. I would also appreciate citations and/or documentation on the relationship between QUBO logical variables (especially for different types of QUBOS, if the specific QUBO matters) and number of qubits in a quantum annealer.