‘Obvious gaps’ exist in the code-writing and operative abilities between domestic and first-class foreign models, an assessment by Tsinghua University found.