ArticleBMC oral health2026
Large language models' performances regarding common patient questions about broken endodontic instruments: a comparative analysis of ChatGPT-5.2, Gemini 3 and DeepSeek V3.2 in accuracy, consistency and readability.
Article in BMC oral health, 2026. The graph could read no effect estimate from its abstract, so it casts no vote on the map. Cited by 1 paper.
What it found
Each row is one number read from the abstract, on the scale the paper reported it, with its interval. Left of the dashed line favours the treatment, right favours the comparator. Under each row is the sentence it came from. New to these charts? A ten-minute tutorial.
The abstract states no effect estimate the extractor could read, or names no intervention and outcome on the map, so this paper lights no cell and moves no belief. It is still indexed, cited and linked below.
The trial behind it
Trials whose registry record cites this paper, or whose number appears in the abstract. A trial that started after this paper was published is citing it as background, not reporting it.
Neither the registry nor the abstract names a trial number. If this is a trial report, that itself is worth knowing.
Who cites it
1 citing paper in PubMed.
Corrections and comments
PubMed lists nothing against this paper. Absence here is not a guarantee, only a check that was made.
Authors and funding
4 authors.
Funding
No grant is acknowledged in the PubMed record.
Abstract
objectiveThis study aims to evaluate and compare the performance of patient education materials generated by five widely used chatbots, ChatGPT-5.2, ChatGPT-5.2 Plus, Gemini 3.0, Gemini 3.0 Plus and DeepSeek V3.2, on answering questions related to broken endodontic instruments in root canals. MATERIALS AND
methodsTwenty-two questions were formulated by two endodontists, each with eight years of experience in instrument removal procedures, based on their clinical expertise and educational materials from the American Association of Endodontists (AAE). The questions were posed to the chatbots over a period of five days, at three different times each day (morning, afternoon, and evening). Two blinded evaluators independently assessed responses for accuracy using a 1-5 scale. Disagreements on scoring were resolved through evidence-based discussions. Coefficient of variation (CV) was calculated to evaluate the consistency of repeated responses for each chatbot. Readability was evaluated using the Flesch Kincaid Reading Ease Score, Flesch Kincaid Grade Level, and SMOG Indices.
resultsSignificant differences in accuracy were found among the chatbots (p < 0.05), with ChatGPT-5.2 demonstrating lower accuracy than the other models (p < 0.001). Accuracy was higher on day 2 than on the other days (p < 0.001). Consistency scores differed significantly among models (p < 0.05), with Gemini 3 Plus and DeepSeek V3.2 showing higher consistency than ChatGPT-5.2. Readability analysis indicated that ChatGPT-5.2 Plus generated more readable responses, whereas Gemini and DeepSeek V3.2 required higher reading grade levels.
conclusionLarge language models (LLMs)-based chatbots showed model-dependent differences in accuracy, consistency, and readability. While Gemini 3, Gemini 3 Plus and DeepSeek V3.2 performed better in terms of accuracy and consistency, ChatGPT-5.2 and ChatGPT-5.2 Plus provided more readable content, highlighting the need for cautious and selective use of these tools in patient education.
Indexed as
Identifiers
What OpenQuestion holds
Registered trials
Read under generation 80e0d062 · epoch 390. Bibliography from PubMed, PubMed Central and OpenAlex; grants from NIH RePORTER; trial links from ClinicalTrials.gov; estimates, votes and beliefs from the OpenQuestion graph.