# Open-ended Health Reasoning on HealthBench

Source: [Open-ended Health Reasoning on HealthBench](<https://www.sota2.com/research/sota/open-ended-health-reasoning-on-healthbench>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-05-25T08:37:33.322Z

Evaluation of model performance on the Open-ended Health Reasoning task using the HealthBench dataset.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Open-ended Health Reasoning |
| Dataset (including any stated split) | HealthBench |
| Metric | Score |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 5 of 5 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [Score](<https://www.sota2.com/research/sota/open-ended-health-reasoning-on-healthbench.md?metric=Score>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Score | Result date | Report paper |
| --- | --- | --- | --- |
| ARES-RL | 41.45 | Not recorded | 6a13d7656f71c5ae7dc9e771 |
| Webscale | 36.08 | Not recorded | 6a13d7656f71c5ae7dc9e771 |
| ARES-SFT | 35.78 | Not recorded | 6a13d7656f71c5ae7dc9e771 |
| CPT | 35.04 | Not recorded | 6a13d7656f71c5ae7dc9e771 |
| Natural Reasoning | 32.94 | Not recorded | 6a13d7656f71c5ae7dc9e771 |

### ARES-RL

Score: 41.45. Result ID: 6a140a07fbc373592c0b3a82.

[Reporting paper](<https://www.sota2.com/research/paper/6a13d7656f71c5ae7dc9e771.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a13d7656f71c5ae7dc9e771.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Webscale

Score: 36.08. Result ID: 6a140a07fbc373592c0b3a80.

[Reporting paper](<https://www.sota2.com/research/paper/6a13d7656f71c5ae7dc9e771.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bec1e209131cea8b4e8.md>)

[Implementation](<https://github.com/SalesforceAIResearch/PretrainRL-pipeline>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### ARES-SFT

Score: 35.78. Result ID: 6a140a07fbc373592c0b3a81.

[Reporting paper](<https://www.sota2.com/research/paper/6a13d7656f71c5ae7dc9e771.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a13d7656f71c5ae7dc9e771.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### CPT

Score: 35.04. Result ID: 6a140a07fbc373592c0b3a7e.

[Reporting paper](<https://www.sota2.com/research/paper/6a13d7656f71c5ae7dc9e771.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Natural Reasoning

Score: 32.94. Result ID: 6a140a07fbc373592c0b3a7f.

[Reporting paper](<https://www.sota2.com/research/paper/6a13d7656f71c5ae7dc9e771.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79721e209131cea7da8f.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.