The accuracy and repeatability of OpenEvidence on complex medical subspecialty scenarios: a pilot study
Bibliographic record
Abstract
Abstract OpenEvidence is a popular artificial intelligence (AI) based medical search engine that generates evidence-based answers. It includes a quick search engine method (OE) that takes only seconds to respond, along with a limited number of references. In mid-2025, the platform introduced “Deep Consult” (DC), which takes several minutes to respond and provides more comprehensive answers with additional references. OpenEvidence scored 100% on USMLE-type multiple-choice questions, but it has not been tested on more complex medical scenarios. We tested the OE and DC models using questions primarily derived from medical specialty board exams, specifically, the MedXpertQA dataset. In a prior published study, this dataset was evaluated with eleven large language models (LLMs), and the results indicated poor accuracy (14-46%) for all LLMs. We evaluated the performance of OpenEvidence on a sample of the MedXpertQA dataset, comprising 100 medical subspecialty scenarios and using two independent evaluators. The highest accuracy for DC was 41%, and for OE, 34%. Repeatability testing revealed an evaluator concordance rate of 77% for OE and 72% for DC.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.021 | 0.120 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".