Typically, is not a numerical system, but rather a scale that changes from PI's and students. You will see this difference in collaboration papers.
The first authors are usually the graduate students listed in their order of importance to the work of the paper (ie the first student listed contributed the most, the second contributed second most, etc). Then come the post-docs and PIs. These are usually posted in the reverse order of the students associated with the collaboration.
Example:
School A has Student A, Post-Doc A, and PI A.
School B has Student B1, Student B2, Post-doc B, and PI B.
The paper is mostly Student A's work, but all 7 parties are involved. The authorship would likely go:
Student A, Student B1, Student B2, Post-Doc B, Post-Doc A, PI B, PI A.
Now let's say Student B2 did the bulk of the work:
Student B2, Student B1, Student A, Post-Doc A, Post-Doc B, PI A, PI B.
The only real variance you will see is with the post-docs. Some will be listed in with the students, some will be in the middle in reverse importance like the professors, some will be in the middle in order of importance like the students.
As a final note, traditionally the "last" author isn't usually the highest honor on a paper, but rather the corresponding author (who is almost always the last author).