What this selection shows

Where MontMedical Pro Model is stronger—and where it needs work.

MontMedical Pro Model has the highest mean in this five-profile selection, at 9.48/10. This describes the selected answers and tested settings; it does not establish broad clinical superiority.

The clearest advantage over OpenEvidence is precise interpretation of what the evidence establishes. Examples include preparation-level versus molecule-level evidence in the prothrombin case, both corrected flow/pressure inputs in the atrial-pressure case, and published variant association versus patient-cell assignment in the JAK3 case. MontMedical Pro scores higher than OpenEvidence on 44 shared cases, lower on 2, with 2 ties.

OpenEvidence’s weaknesses here are mainly over-specific causal or clonal conclusions and incomplete interpretation of specialised testing. Many answers still identify the leading disease correctly. Its stronger answers remain visible: OpenEvidence outscored this MontMedical Pro run on NCF4/p40phox deficiency and riboflavin transporter deficiency.

Fable 5.2 Max combines strong disease recognition with some overconfident experimental conclusions. Its shared-case mean is 8.94/10, and its mean over all 49 answers is 8.96/10. Strong examples include masked polycythaemia vera, glucocorticoid resistance and riboflavin-transporter deficiency. Larger deductions concern unproved amyloid typing, SAMD9L phase and clonal history, and pulmonary haemodynamic interpretation. Highlighted passages and explanatory notes show the specific concerns; three Provider errors remain ungraded.

DeepSeek Flash Max and Carbon High recognise many of the rare disorders. Their more frequent deductions concern treating plausible mechanisms as established facts—for example, amyloid precursor typing, clone identity, and assay signals that do not directly measure the claimed biological effect. Every case and its rating note is available above.

These findings describe the selected answers and tested profiles. Independent, blinded evaluation on new cases is needed to establish clinical diagnostic accuracy or patient benefit.