Med42-v2-8B: benchmark results

Provider: Other.

Aggregate Rating 1487 ± 43, rank #2558 of 10916 rated models, from 13 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
MR-Bench41.7Accuracy (%, times 100), mean of the medication-imputation a65.2
MR-Bench - Medication Imputation24.9Accuracy (%, times 100) at picking the masked medication amo65.2
MR-Bench - Procedure Selection58.5Accuracy (%, times 100) at picking the admission's procedure56.5
MINT (Medical) - MedQA (Question First, Initial Answer)38.7Initial-answer accuracy (%) on the 600 MedQA-derived diagnos44.4
MINT (Medical) - MedMCQA (Question Last)79.9Final-answer accuracy (%) on the 174 MedMCQA-derived diagnos40
MINT (Medical) - MedQA (Question Last)62.5Final-answer accuracy (%) on the 600 MedQA-derived diagnosis40
MINT (Medical) - MedBullets (Question First, Initial Answer)29Initial-answer accuracy (%) on the 62 MedBullets-derived dia33.3
MINT (Medical) - Derm-Public (Question Last)75Final-answer accuracy (%) on the 100 Derm-Public-derived dia30
MINT (Medical) - MedBullets (Question Last)54.8Final-answer accuracy (%) on the 62 MedBullets-derived diagn30
MINT (Medical) - MedMCQA (Question First, Initial Answer)50Initial-answer accuracy (%) on the 174 MedMCQA-derived diagn30
MINT (Medical) - Derm-Private (Question First, Initial Answer)33.7Initial-answer accuracy (%) on the 99 Derm-Private-derived d20
MINT (Medical) - Derm-Private (Question Last)69.7Final-answer accuracy (%) on the 99 Derm-Private-derived dia20

Interactive version: theaggregate.ai/model?slug=med42-v2-8b · How It Works · Data refreshed daily, snapshot 2026-11-08.