MedCTA: leaderboard
MedCTA evaluates medical tool agents on clinician-validated, step-implicit tasks grounded in multimodal clinical inputs, including radiology images, pathology slides, and reports. The benchmark contains 107 real-world clinical tasks with clinician-verified executable trajectories over five deployed tools and measures tool selection, argument validity, execution stability, trajectory fidelity, and
Metric: Outcome Accuracy (self-reported). Source: benchmarklist.com. Saturation forecast: Around July 2027. 18 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.4 | 31.54 | #79 |
| 2 | Claude Opus 4.6 | 31.32 | #63 |
| 3 | GPT-5.4 Mini | 28.31 | #212 |
| 4 | Qwen 3 8B | 27.8 | #674 |
| 5 | Gemini 3 Flash | 25.87 | #99 |
| 6 | Claude Sonnet 4.6 | 25.33 | #90 |
| 7 | Claude Haiku 4.5 | 23.08 | #280 |
| 8 | Qwen 3.5 9B | 21.64 | #379 |
| 9 | GPT-5.4 Nano | 20.3 | #328 |
| 10 | Llama 3.1 8B Instruct | 18.94 | #1035 |
| 11 | Llama 3.2 3B Instruct | 11.29 | #1352 |
| 12 | deepseek-llm-7B-chat | 11 | #1354 |
| 13 | Phi-4 | 10.65 | #706 |
| 14 | Mistral 7B | 9.4 | #1397 |
| 15 | DeepSeek-V2-Lite-Chat | 6.96 | #1210 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=medcta · How It Works · Data refreshed daily, snapshot 2026-10-20.