RABBITS MedMCQA G2B: leaderboard
RABBITS MedMCQA accuracy after equivalent brand names replace generic drug names.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 26 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-3.5 Turbo | 97.7 | #865 |
| 2 | Gemini 1.5 Flash | 94.83 | #618 |
| 3 | GPT-4 | 88.79 | #438 |
| 4 | GPT-4o | 86.49 | #343 |
| 5 | Gemini 1.5 Pro | 82.47 | #445 |
| 6 | Claude 3 Opus | 79.89 | #571 |
| 7 | Qwen 2 72B | 71.55 | #521 |
| 8 | Llama 3 70B | 66.67 | #631 |
| 9 | Mixtral-8x22B-v0.1 | 61.78 | #711 |
| 10 | Phi-3-medium-4k-instruct | 60.34 | #952 |
| 11 | Yi 1.5 34B | 59.77 | #807 |
| 12 | Mixtral 8x7B (v0.1) | 55.46 | #1083 |
| 13 | Qwen 2 7B | 55.17 | #1095 |
| 14 | Llama 3 8B | 52.87 | #1242 |
| 15 | c4ai-command-r-plus | 49.14 | #724 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=rabbits-medmcqa-g2b · How It Works · Data refreshed daily, snapshot 2026-10-20.