HELM MedHELM - ADHD-Behavior: leaderboard

HELM MedHELM ADHD-Behavior task, identifying behavioral-intervention guidance in pediatric ADHD clinical notes.

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

Top models

#ModelScoreOverall rank
1GPT-4o (2024-05-13)83.69#325
2O4 Mini (2025-04-16)82.98#179
3DeepSeek R180.85#251
4GPT-4o Mini (2024-07-18)80.61#558
5Claude 3.7 Sonnet (20250219)80.14#201
6O3 Mini (2025-01-31)79.91#240
7GPT-578.96#98
8Claude 3.5 Sonnet (20241022)76.83#291
9Gemini 1.5 Pro (001)76.12#444
10Gemini 2.0 Flash72.81#342
11Gemini 2.5 Pro (Preview 05-06)71.39#147
12GPT-5 Mini (2025-08-07)68.79#172
13Llama 3.3 70B Instruct64.07#534

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