BRIDGE Medical Leaderboard: leaderboard
Clinical text understanding across 87 tasks. This card reports the published zero-shot average; five-shot and chain-of-thought evaluations are separate boards.
Metric: Average Performance (%). Source: huggingface.co. Source: huggingface.co. 112 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemma 4 31B (IT) | 46.74 | #176 |
| 2 | Gemini 2.5 Flash | 44.84 | #245 |
| 3 | Qwen 3.8 27B (Non-reasoning) | 44.6 | #105 (Qwen 3.8 27B) |
| 4 | Gemma 4 26B A4B (IT) | 44.51 | #297 |
| 5 | DeepSeek R1 | 44.25 | #251 |
| 6 | GPT-4o (2024-08-06) | 44.2 | #336 |
| 7 | Qwen 3 Next 80B A3B (Thinking) | 43.86 | #320 (Qwen 3 Next 80B A3B) |
| 8 | Gemini 1.5 Pro (002) | 43.85 | #349 |
| 9 | Gemini 2.0 Flash (001) | 43.03 | #360 |
| 10 | Mistral Large 2 (Nov) Instruct (2411) | 42.28 | #465 |
| 11 | Qwen 3 30B A3B 2507 (Thinking) | 41.82 | #376 (Qwen 3 30B A3B 2507) |
| 12 | Qwen 3 235B A22B (Thinking) | 41.63 | #313 (Qwen 3 235B A22B) |
| 13 | Qwen 2.5 72B Instruct | 41.62 | #454 |
| 14 | Qwen 3 32B (Thinking) | 41.04 | #443 (Qwen 3 32B) |
| 15 | Qwen 3 30B A3B (Thinking) | 40.93 | #493 (Qwen 3 30B A3B) |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=bridge-medical-leaderboard · How It Works · Data refreshed daily, snapshot 2026-10-20.