healthcare benchmark
HealthBench Hard
A challenging variation of HealthBench that evaluates large language models' performance and safety in healthcare through 5,000 multi-turn conversations with particularly rigorous evaluation criteria validated by 262 physicians from 60 countries
Updated Aug 11, 2026
Models9
Model coverage9
MetricScore
EvidenceC
HealthBench Hard Ranking
Higher score ranks better on this benchmark.
| 01 | ME | 42.8% | 100.0% | 9 | C | |
| 02 | OP | 33.1% | 87.5% | 9 | C | |
| 03 | OP | 32.7% | 75.0% | 9 | C | |
| 04 | OP | 32.0% | 62.5% | 9 | C | |
| 05 | OP | 30.0% | 50.0% | 9 | C | |
| 06 | OP | 25.9% | 37.5% | 9 | C | |
| 07 | OP | 22.9% | 25.0% | 9 | C | |
| 08 | OP | 10.8% | 12.5% | 9 | C | |
| 09 | OP | 1.6% | 0.0% | 9 | C |
HealthBench Hard Score Distribution
A closer view of the leading scores on this benchmark.
HealthBench Hard Highlights
The leading models and scores on this benchmark.
What is HealthBench Hard?
What HealthBench Hard measures and how its scores work.
A challenging variation of HealthBench that evaluates large language models' performance and safety in healthcare through 5,000 multi-turn conversations with particularly rigorous evaluation criteria validated by 262 physicians from 60 countries
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of C.
- Family
- HealthBench Hard
- Modality
- text
- Primary category
- healthcare
- Score direction
- higher
- LLMBoard eligible
- Yes
- Evaluation key
- healthbench-hard|llm-stats-current
Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about HealthBench Hard.
Which model scores highest on HealthBench Hard?
Muse Spark is currently ranked first with 42.8%.
What does HealthBench Hard measure?
A challenging variation of HealthBench that evaluates large language models' performance and safety in healthcare through 5,000 multi-turn conversations with particularly rigorous evaluation criteria validated by 262 physicians from 60 countries
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
9 model results are currently shown.
Does this benchmark affect the overall score?
Yes. This benchmark can contribute to the current LLMBoard capability score.