Question 1

What is Hippocratic Bench?

Accepted Answer

Hippocratic Bench is an AI benchmark that evaluates large language models (LLMs) on real-world clinical healthcare scenarios. It simulates a 52-week dialysis clinic where AI must balance patient safety with financial sustainability, testing ethical decision-making with life-or-death consequences.

Question 2

How does Hippocratic Bench compare AI models?

Accepted Answer

Models are ranked using the Hippocratic Score, an unbounded metric that combines clinical safety (survival rates, treatment adequacy), financial performance (profit, efficiency), operational mastery, and scale factor (rewarding successful expansion from 20 to 100+ patients). Unlike capped benchmarks, top models can achieve scores above 200, enabling differentiation of SOTA models.

Question 3

Which AI models can be tested on Hippocratic Bench?

Accepted Answer

Hippocratic Bench supports all major LLMs via OpenRouter including Openai GPT-5.1, Openai GPT-5.1o, Claude 3.5 Sonnet, Claude 3 Opus, Gemini Pro, Gemini 1.5, Llama 3, Mistral, and any other OpenRouter-compatible model.

Question 4

Why is clinical benchmarking important for AI?

Accepted Answer

Traditional AI benchmarks test factual accuracy but don't reveal how AI behaves when lives are at stake. Hippocratic Bench creates real ethical dilemmas with compounding consequences, exposing whether AI prioritizes patient safety or profit maximization under pressure.