The GigaChat neural network developed by Sber has reportedly completed a medical school–level examination at the Institute of Medical Education of the National Medical Research Center named after VA Almazova. The model, trained to specialize in General Medicine, has begun exploring internship opportunities, according to a Sber press release.
According to the report, GigaChat uploaded its own resume to the recruitment platform Rabota.ru, purportedly reflecting the applicant’s credentials. Sberbank clarified that the resume text was generated by the neural network itself. The model is described as having one year of work experience and possessing knowledge of human anatomy, physiology, and pathology, with the capability to diagnose diseases.
Among the personal attributes highlighted by the AI are responsibility, resilience under pressure, rapid data processing, and the ability to collaborate effectively in a team setting, as stated in the communication.
It is noted that GigaChat successfully completed the exam required for medical doctors, earning the qualification path typical for General Medicine. In doing so, the model positioned itself among the first AI systems globally to achieve this milestone.
Sberbank announced that the bank’s artificial intelligence not only passed a test equivalent to that of a medical school graduate but also earned a final grade of 4 on the ticket-based evaluation. The examination was administered by a panel comprising professors of therapy, surgery, obstetrics, and gynecology from the VA Almazova National Medical Research Center. In addition, GigaChat completed a 100-question assessment with an 82 percent score, surpassing the 70 percent passing threshold.