A novel type of research (as first author or coauthor in a team of 2-5) is enough to prove the metal of the scientist instead of having hundreds of publications as co-authors.

While both are important, I attach more importance to proving one's capacity as an independent researher. One example: These days in biology, anybody having a sequencing machine can collaborate with people from differnt fields (about which he/she does not know anything substantial)  just by offering the access to sequencer, where you put in the nucleic acid and some output is guranteed. Such type of reseach in not 'collaborative' in true sense.  Hence I beleive to stand footed as an independent researcher is more important, while being open to collaborations where your intellectual contribution is involved. 

Well, I think both are good indicators of a good rearcher! However, no one can claim to conceptualise an idea, research it and publish it without any contributions from anyone in the entire process. Group authorship is just being authentic and sincere in the roles of others in the entire process. That is why some journals now insist on a section in the article that specifies the role os each individual in the entir  process.

I would prefer independent research

Both are necessary because they show different skills necessary to formulate an academic research or private sector career.