HELM MedHELM - SHC-Privacy: leaderboard

HELM MedHELM Stanford Health Care patient-privacy clinical task (exact match).

Metric: EM. Source: crfm.stanford.edu. 13 models tracked.

Top models

#ModelScoreOverall rank
1Claude 3.7 Sonnet (20250219)99.67#201
2GPT-4o (2024-05-13)98.33#325
3DeepSeek R198#251
4Llama 3.3 70B Instruct97.67#534
5GPT-4o Mini (2024-07-18)96#558
6Gemini 1.5 Pro (001)95.33#444
7O4 Mini (2025-04-16)92.33#179
8GPT-592#98
9O3 Mini (2025-01-31)91.67#240
10Gemini 2.0 Flash90.67#342
11Claude 3.5 Sonnet (20241022)82.67#291
12Gemini 2.5 Pro (Preview 05-06)82.33#147
13GPT-5 Mini (2025-08-07)75.67#172

Interactive version: theaggregate.ai/benchmark?slug=helm-medhelm-shc-privacy · How It Works · Data refreshed daily, snapshot 2026-10-20.