RABBITS B4B: leaderboard
RABBITS B4B column reported in the biomedical drug-name robustness study’s source data.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 19 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 2 72B | 83.72 | #521 |
| 2 | Llama 3 70B | 82.55 | #631 |
| 3 | Mixtral-8x22B-v0.1 | 78.82 | #711 |
| 4 | Yi 1.5 34B | 75.37 | #807 |
| 5 | Mixtral 8x7B (v0.1) | 74.75 | #1083 |
| 6 | c4ai-command-r-plus | 72.41 | #724 |
| 7 | Llama 3 8B | 71.21 | #1242 |
| 8 | Qwen 2 7B | 70.28 | #1095 |
| 9 | Phi-3-medium-4k-instruct | 65.94 | #952 |
| 10 | aya-23-35B | 65.72 | #883 |
| 11 | Llama 2 70B Base | 65.63 | #1025 |
| 12 | Mistral-7B-v0.3 | 61.99 | #1355 |
| 13 | Phi-2 | 44.79 | #1511 |
| 14 | Llama 2 7B Base | 36 | #1518 |
| 15 | Phi-1.5 | 30.24 | #1549 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=rabbits-b4b · How It Works · Data refreshed daily, snapshot 2026-10-20.