MedAraBench: leaderboard
Metric: Overall Accuracy (%). Source: benchmarklist.com. 16 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | O3 | 76.5 |
| 2 | GPT-5 | 76.4 |
| 3 | Claude Sonnet 4 (20250514) | 69.4 |
| 4 | GPT-4.1 | 67.3 |
| 5 | Gemini 2.0 Flash | 65.4 |
| 6 | DeepSeek V3 (0324) | 62 |
| 7 | Qwen Plus | 61.8 |
| 8 | Llama 3.3 70B Instruct | 54.7 |
| 9 | MedGemma-4B-IT | 39 |
| 10 | c4ai-command-r7B-arabic-02-2025 | 38.1 |
| 11 | apollo-7B | 23.8 |
| 12 | Llama 3.1 8B Instruct | 17 |
Interactive version: theaggregate.ai/benchmark?slug=medarabench · How It Works · Data refreshed daily, snapshot 2026-09-19.