MedReason-8B: benchmark results
Provider: Other.
Aggregate Rating 1432 ± 31, rank #4586 of 10916 rated models, from 18 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| HiMed-West - MMLU-Pro Health (Hindi) | 27.6 | Accuracy (%; 818 translated MMLU-Pro Health questions, zero- | 75 |
| HiMed (Hindi) | 37.9 | Accuracy (%; mean of the five Hindi sets, zero-shot multiple | 60 |
| HiMed-West - MMLU-Pro Biology (Hindi) | 43.7 | Accuracy (%; 717 translated MMLU-Pro Biology questions, zero | 60 |
| CureMed-Bench - Language Consistency | 26.74 | Language consistency (%; share of final answers written in t | 47.9 |
| CureMed-Bench | 34.12 | Logical accuracy (%; share of final answers a GPT-4o judge m | 43.8 |
| HiMed (English) | 51.3 | Accuracy (%; mean of the four English sets, zero-shot multip | 40 |
| HiMed-Trad | 47.2 | Accuracy (%; 6,010 Hindi questions on Indian systems of medi | 40 |
| HiMed-West Exam (Hindi) | 41.3 | Accuracy (%; 470 NEET-UG questions in Hindi, zero-shot multi | 40 |
| MR-Bench - Medication Imputation | 22 | Accuracy (%, times 100) at picking the masked medication amo | 32.6 |
| HiMed-West Exam (English) | 53.2 | Accuracy (%; 470 NEET-UG questions in English, zero-shot mul | 30 |
| MedAgentGym | 18.16 | Average Score (self-reported) | 25 |
| MR-Bench | 35.1 | Accuracy (%, times 100), mean of the medication-imputation a | 17.4 |
Interactive version: theaggregate.ai/model?slug=medreason-8b · How It Works · Data refreshed daily, snapshot 2026-11-08.