index / FeiLiuEM/open-medical-jev
open-medical-jev
Jev-level results from frozen open models: within ~2 pts of Jev on national medical exams (≈3× Laya; level with OpenJev). No fine-tuning, no distillation, no corpus and high compatibility with new models. Three modes: fast runs at ≈0.076s per question; general and high cascades cut compute ≈63%/49% at 93%/97% released precision.
Research and EvalsVerificationListedPython
Repository
Stars21
LanguagePython
LicenseMIT
Created2026-09-25
Last pushed2026-09-27
Officialno
Judgment
Genuine Jev project0.71
gate 0.5
Uses Jev at runtime0.10
Is a meta list0.04
Is a reimplementation0.90
Quality scales
Substance2.61 / 3
gate 0.5
Docs2.98 / 3
Novelty2.14 / 3
Composite score0.90
Category distribution
Research and Evals1.00
Integrations0.00
SDKs and Clients0.00
Agent and Dev Tooling0.00
Learning0.00
Other0.00
Applications0.00
Games and Simulation0.00
Assigned categoryResearch and Evals
Category probability1.00
Category confidence0.99
PatternVerification
Pattern confidence0.33
Curation
StatusListed
Reasongate passed
README pickno
Overridenone
Sources
- search:topic:jev
- search:jev in:name,description,readme created:2026-09-25
- list:AbdelStark/awesome-typesafe
Provenance
Modeljev-1.13.0
Question setv2
Judged at2026-09-27 11:42 UTC