In general the answer is NO. But, clearly the number of co-authors is depend on the paper complexity and type of the paper (regular paper, survey paper, ...).
Usually conference papers are two (student and advisor) or three (principle student, helping student, and the advisor), but for journal papers it generally ranges from 2 to 6 or even 8, but I don't think more is normal.
No way that one can set a limit, but there is a way to limit the number using a quantitative measure of author's contribution for the production of a manuscript. That is termed as "relative intellectual contribution" of an author. You may like to see the paper below:
Article The need to quantify authors’ relative intellectual contribu...
A research paper can have two to five collaborating authors, and this is considered a normal collaborating group. From six co-authors and up this is already a large team work. It depends on the discipline, in the social sciences there are usually 2 to 4 co-authors per paper, and in the sciences there are often many more . I have seen papers in physics produced by 200 co-authors.
it depends on the field you are working in, in electrical engineering I have seen ten co-authors per paper. Normally, in engineering up ten co-authors are acceptable.
As others have mentioned, I have seen numerous co-authors sometimes more than words in the abstract! One colleague did tell me that the rule of thumb is that the more "important" contributors to the work are the first author, middle and the last author. First is the one who does the majority of the bench work and writing, last author is the P.I. who foots the bill with the grant and the middle is the person who does the second most physical work on the data collection. The rest are ones who contribute in other ways, donate antibodies, contribute significant ideas or help with data analysis.