QuarkMedBench - Important: leaderboard
Metric: Important score rate (%): share of the Important rubric points (logical completeness and explanatory depth) that the answer earns, on a stratified sample of 2,000 real Chinese online medical and health queries (single- and multi-turn), answers scored against query-specific rubrics (Essential, Important, Highlight and Pitfall items) by a Qwen3-235B judge, without an output length limit; higher is better. Source: arxiv.org. Saturation forecast: Around September 2027. 13 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Pro | 74.64 | #145 |
| 2 | Qwen 3 30B A3B 2507 (Thinking) | 74.37 | #366 (Qwen 3 30B A3B 2507) |
| 3 | GPT-5.1 | 72.32 | #131 |
| 4 | DeepSeek V3.2 | 72.29 | #198 |
| 5 | DeepSeek R1 0528 | 70.26 | #217 |
| 6 | Qwen 3 235B A22B (Non-reasoning) | 69.67 | #304 (Qwen 3 235B A22B) |
| 7 | Gemini 3 Pro | 68.64 | #77 |
| 8 | Qwen 3 32B (Non-reasoning) | 68.48 | #424 (Qwen 3 32B) |
| 9 | GPT-5.2 | 65.39 | #105 |
| 10 | Seed-1.6 | 62.31 | #257 |
| 11 | Qwen 3 30B A3B (Non-reasoning) | 60.87 | #488 (Qwen 3 30B A3B) |
| 12 | Qwen 3 8B (Non-reasoning) | 60.55 | #667 (Qwen 3 8B) |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=quarkmedbench-important · How It Works · Data refreshed daily, snapshot 2026-10-11.