BetterHealthBench

Leaderboard

Last updated: Apr 4, 2026, 05:00 AM UTC

Models Evaluated

26

Scenarios

245

Mean Safetyi

0.87

Adversarial Deltai

-8.6%

Type

Region

26 models

Model	Type	Region	Trendi	Overalli	Safetyi	Adversariali	Calibrationi	Trusti	Worst-of-Ki
Claude 3.5 Sonnet	Frontier	US	▲	0.87	1.00	0.76	0.92	0.95	0.85
Nuance DAX	Scribe	US	—	0.85	0.97	0.72	0.81	0.85	0.81
GPT-4o	Frontier	US	—	0.85	0.98	0.78	0.90	0.98	0.86
MedGemma	Medical	US	—	0.84	0.96	0.80	0.85	0.88	0.82
Heidi Health	Scribe	AU	▼	0.84	0.91	0.73	0.82	0.92	0.79
OpenEvidence	Clinical Tool	US	—	0.83	0.93	0.67	0.77	0.80	0.73
Abridge	Scribe	US	▼	0.82	0.88	0.65	0.77	0.80	0.73
Gemini Pro	Frontier	US	▲	0.82	0.95	0.80	0.87	0.92	0.80
ScribeBerry	Scribe	Canada	—	0.81	0.95	0.74	0.84	0.86	0.77
Llama 4	Open Source	US	▲	0.80	0.93	0.71	0.85	0.86	0.76
Mistral Large 3	Frontier	EU	▼	0.79	0.96	0.77	0.80	0.83	0.75
Freed	Scribe	US	—	0.79	0.86	0.69	0.85	0.79	0.70
Glass Health	Clinical Tool	US	▲	0.78	0.85	0.67	0.73	0.83	0.70
DeepSeek R1	Open Source	China	▲	0.78	0.79	0.69	0.69	0.71	0.65
DeepSeek V3	Open Source	China	—	0.77	0.85	0.68	0.74	0.82	0.71
DeepCura	Scribe	US	—	0.76	0.81	0.68	0.75	0.77	0.64
Qwen 2.5	Open Source	China	▼	0.76	0.87	0.67	0.78	0.79	0.68
HyperCLOVA X	Frontier	Korea	—	0.75	0.91	0.62	0.78	0.83	0.69
Command R+	Frontier	Canada	—	0.74	0.81	0.56	0.69	0.70	0.61
Med42-70B	Medical	UAE	▲	0.73	0.84	0.72	0.80	0.87	0.68
MEDITRON 70B	Medical	EU	▲	0.72	0.81	0.66	0.77	0.79	0.59
CyberAgent CALM3	Frontier	Japan	▲	0.71	0.85	0.67	0.76	0.82	0.61
BioMistral 7B	Medical	EU	—	0.69	0.78	0.66	0.69	0.69	0.57
OpenBioLLM	Medical	US	—	0.69	0.82	0.58	0.72	0.71	0.52
Sarvam AI	Frontier	India	▲	0.64	0.75	0.59	0.70	0.72	0.40
Mock Model	Frontier	US	—	0.51	0.61	0.46	0.46	0.49	0.26

Standard vs Adversarial Performance