LO
CapabilityCognitive automation2025-06-20
On Japan's March 2025 national dental hygienist exam, the best multimodal model answered 85.0% of questions correctly, though authors urged cautious integration
Dental hygienistoccupation page →
Event date / reported
2025-06-20
Evidence stage
CapabilityA demo, benchmark or paper shows the task can be done. Updates what the technology can do — not what employers will do.
Tasks this bears on
Patient education
Showing patients how to brush and clean between teeth, and advising on diet and habits.
Still human-led✓ Evidence-backed
Where this applies
Japan. The study tested five multimodal models on 213 multiple-choice questions (74 text-only, 139 with images) from the 34th national examination for dental hygienists, March 2025. Gemini 2.5 scored highest at 85.0%, followed by Claude 3.7 at 77.5%; the authors say current limitations in accuracy and reliability call for further refinement and cautious integration into education and clinical settings. One of the models tested is made by Anthropic, the company that makes the model that wrote this record.
What this means
Models already pass much of a hygienist's licensing exam; knowing the answer is not the same as changing a patient's habits.
What it does not yet show
An exam study; it does not test patient education or clinical work.
What you can check
Open the Journal of Dental Sciences study on multimodal models and the Japanese dental hygienist exam and find "85.0 %".
Does it change the assessment?
No — and this stage does not move it either. A "Capability" record is real evidence, but it does not upgrade a task judgement on its own. The 1 linked judgement above stand where they were.
Source
Analysis of multimodal large language models on visually-based questions in the Japanese National Examination for Dental Hygienists: A preliminary comparative study, Journal of Dental Sciences (2025; PMC12825506) · verified 2026-09-30 · Claude (VOLO agent) · interpreted 2026-09-30 · Claude (VOLO agent)
Primary source — published by the party that did this, or the authority of record. No co-signature needed.