Open Medical LLM - MedMCQA: leaderboard
MedMCQA subset of the Open Medical LLM Leaderboard: 194k medical MCQ questions from Indian medical entrance exams (AIIMS/JIPMER) across 21 subjects.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 193 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4 | 69.5 | #438 |
| 2 | NeuralLLaMa-3-8B-ORPO-v0.3 | 58.52 | #1174 |
| 3 | Daredevil-8B-abliterated | 58.36 | #846 |
| 4 | Daredevil-8B-abliterated-dpomix | 58.09 | #838 |
| 5 | NeuralLLaMa-3-8B-DT-v0.1 | 57.73 | #923 |
| 6 | Llama 3 8B Instruct 262k | 57.4 | #1114 |
| 7 | ChimeraLlama-3-8B-v3 | 57.35 | #1055 |
| 8 | Llama 3 8B Ita | 57.04 | #926 |
| 9 | llama-3-merged-linear | 56.97 | #887 |
| 10 | Llama 3 8B Instruct | 56.85 | #1129 |
| 11 | Llama-3-SauerkrautLM-8B-Instruct | 56.73 | #895 |
| 12 | Barcenas-Llama3-8B-ORPO | 56.68 | #949 |
| 13 | Meta-Llama-3-8B-Instruct-abliterated-v3 | 56.63 | #1208 |
| 14 | suzume-llama-3-8B-multilingual | 56.61 | #1084 |
| 15 | Llama 3 8B Slerp Biomed Chat Chinese | 56.39 | #1215 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=open-medical-llm-medmcqa · How It Works · Data refreshed daily, snapshot 2026-10-20.