ClinSQL - Medication Management: leaderboard
ClinSQL's Execution Score (a GPT-5 judge checks the executed result table against an expert rubric tree; percent, higher is better) restricted to the test tasks of its medication management scenario: prescriptions, dosing and regimen durations with temporal reasoning. Queries come from chain-of-thought prompting with up to two execution-error refinements; Yale University and University of Washington, 2026.
Metric: Execution Score (%). Source: arxiv.org. Saturation forecast: Estimated already saturated. 22 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Pro | 77.38 | #262 |
| 2 | GPT-5 Mini (2025-08-07) | 75.06 | #341 |
| 3 | DeepSeek R1 0528 | 72.87 | #425 |
| 4 | GPT-4.1 | 71.03 | #492 |
| 5 | O4 Mini (2025-04-16) | 70.01 | #362 |
| 6 | Gemini 2.5 Flash | 67.96 | #452 |
| 7 | Grok 4 Fast (Reasoning) | 67.01 | #438 (Grok 4 Fast) |
| 8 | Qwen 3 Coder 480B A35B Instruct | 66.38 | #641 |
| 9 | Qwen 3 235B A22B 2507 FP8 (Thinking) | 62.32 | #431 |
| 10 | DeepSeek V3.1 | 62.13 | #500 |
| 11 | Qwen 3 235B A22B 2507 Instruct | 60.35 | #619 |
| 12 | Mistral Medium 3 | 58.6 | #1079 |
| 13 | GPT-5 Nano | 58.35 | #888 |
| 14 | Qwen 3 Next 80B A3B Instruct | 52.76 | #763 |
| 15 | Llama 4 Maverick Instruct | 47.08 | #1125 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=clinsql-medication-management · How It Works · Data refreshed daily, snapshot 2026-11-12.