LingxiDiagBench (Static): leaderboard
Metric: Overall score (times 100, 0-100): mean of the eleven printed classification metrics (accuracy, macro-F1 and weighted F1 of the 2-class and 4-class tasks; accuracy, top-1 and top-3 accuracy, macro-F1 and weighted F1 of the 12-class task); LingxiDiagBench-Static on the 1,000 test cases of LingxiDiag-16K (synthetic Chinese psychiatric consultation dialogues with EMRs matching the Shanghai Mental Health Center outpatient distribution): zero-shot diagnosis from the complete dialogue; higher is better. Source: arxiv.org. Saturation forecast: Around 2032. 13 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Grok 4.1 Fast | 52.1 | #208 |
| 2 | Claude Haiku 4.5 | 51.6 | #271 |
| 3 | Gemini 3 Flash | 51 | #93 |
| 4 | Qwen 3 32B | 50.6 | #424 |
| 5 | GPT-5 Mini | 50.4 | #176 |
| 6 | DeepSeek V3.2 | 50.1 | #198 |
| 7 | Kimi K2 (Thinking) | 48.4 | #236 (Kimi K2) |
| 8 | GPT-OSS-20B | 47.9 | #499 |
| 9 | Qwen 3 4B | 47.4 | #823 |
| 10 | Qwen 3 8B | 47.3 | #667 |
| 11 | Qwen 3 1.7B | 44.8 | #1186 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=lingxidiagbench-static · How It Works · Data refreshed daily, snapshot 2026-10-11.