We present a case study of proving by three 12-13-year-old students with different levels of mathematical giftedness. After analysing students' proofs, we conclude that: there was a relation on the consistency and the students' levels of mathematically giftedness, being the least consistent the student not mathematically gifted and the most consistent the student with the highest level of mathematical giftedness; the variability was greater in the arithmetical problems; the quality of the proofs produced increased as the level of mathematical giftedness did; the two students with a lower level did better proofs in the arithmetical than in the geometrical problems, while the student with the highest level did not show significant differences between the two areas.