HEART
Humans and models respond to people seeking emotional support. Judges compare responses and explain their preferences.
- GPT-5 1,612
- o3 1,588
- Claude Haiku 4.5 1,562
- Kimi K2 Sep 2025 1,553
- Claude Sonnet 4.5 1,541
Official leaderboard
We track 1 benchmark and 1 arena that Hippocratic AI made or helped make.
Top five in each of its 1 arena with three or more rated models, the most rated first.
Humans and models respond to people seeking emotional support. Judges compare responses and explain their preferences.
Official leaderboard
| Name | Kind | Released | Best score | Status |
|---|---|---|---|---|
| Polaris Healthcare Safety Benchmarks Models conduct simulated healthcare conversations. The evaluation measures clinical safety, communication, and administrative task performance. | Medicine | Not in cards |