SurgiQ - Reasoning: leaderboard
Metric: Reasoning questions (2,673): accuracy (%) on SurgiQ's four-option surgical multiple-choice questions, zero-shot; the answer is the option label with the highest next-token log-probability (plain and space-prefixed labels), options shuffled once. Source: arxiv.org. Saturation forecast: Around December 2026. 35 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 2.5 72B Instruct | 71.25 |
| 2 | Qwen 2.5 72B | 70.83 |
| 3 | GPT-OSS-120B | 67.91 |
| 4 | Qwen 2.5 32B | 67.45 |
| 5 | Qwen 2.5 32B Instruct | 67.15 |
| 6 | Ministral-3-14B-Reasoning-2512 | 65.69 |
| 7 | GPT-OSS-20B | 62.16 |
| 8 | DeepSeek R1 Distill Qwen 32B | 61.56 |
| 9 | K2 Think V2 | 60.21 |
| 10 | Ministral-3-8B-Instruct-2512 | 60.17 |
| 11 | DeepSeek R1 Distill Qwen 14B | 57.81 |
| 12 | BioMistral-7B | 50.53 |
| 13 | MedGemma-4B-IT | 48.61 |
| 14 | Gemma 2 9B (IT) | 44.41 |
| 15 | MedGemma-27B-IT | 40.35 |
Interactive version: theaggregate.ai/benchmark?slug=surgiq-reasoning · How It Works · Data refreshed daily, snapshot 2026-09-29.