# Medical Question Answering on GenMedGPT-5k 1.0 (test)

Source: [Medical Question Answering on GenMedGPT-5k 1.0 (test)](<https://www.sota2.com/research/sota/medical-question-answering-on-genmedgpt-5k-1-0-test>)

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

Data updated: 2026-02-26T05:41:13.825Z

Evaluation of model performance on the GenMedGPT-5k 1.0 test set for Medical Question Answering, assessed using win/tie/loss metrics across factualness, disease diagnosis, and drug recommendation categories.

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 | Medical Question Answering |
| Dataset (including any stated split) | GenMedGPT-5k 1.0 (test) |
| Metric | Diversity & Integrity Win Rate |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 6 of 6 |
| 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

- [Diversity & Integrity Win Rate](<https://www.sota2.com/research/sota/medical-question-answering-on-genmedgpt-5k-1-0-test.md?metric=Diversity%20%26%20Integrity%20Win%20Rate>): Higher is better; unit: PERCENT.

- [Diversity & Integrity Tie Rate](<https://www.sota2.com/research/sota/medical-question-answering-on-genmedgpt-5k-1-0-test.md?metric=Diversity%20%26%20Integrity%20Tie%20Rate>): Lower is better; unit: PERCENT.

- [Diversity & Integrity Loss Rate](<https://www.sota2.com/research/sota/medical-question-answering-on-genmedgpt-5k-1-0-test.md?metric=Diversity%20%26%20Integrity%20Loss%20Rate>): Lower is better; unit: PERCENT.

- [Factualness Win Rate](<https://www.sota2.com/research/sota/medical-question-answering-on-genmedgpt-5k-1-0-test.md?metric=Factualness%20Win%20Rate>): Higher is better; unit: PERCENT.

- [Factualness Tie Rate](<https://www.sota2.com/research/sota/medical-question-answering-on-genmedgpt-5k-1-0-test.md?metric=Factualness%20Tie%20Rate>): Lower is better; unit: PERCENT.

- [Factualness Loss Rate](<https://www.sota2.com/research/sota/medical-question-answering-on-genmedgpt-5k-1-0-test.md?metric=Factualness%20Loss%20Rate>): Lower is better; unit: PERCENT.

- [Disease Diagnosis Win Rate](<https://www.sota2.com/research/sota/medical-question-answering-on-genmedgpt-5k-1-0-test.md?metric=Disease%20Diagnosis%20Win%20Rate>): Higher is better; unit: PERCENT.

- [Disease Diagnosis Tie Rate](<https://www.sota2.com/research/sota/medical-question-answering-on-genmedgpt-5k-1-0-test.md?metric=Disease%20Diagnosis%20Tie%20Rate>): Lower is better; unit: PERCENT.

- [Disease Diagnosis Loss Rate](<https://www.sota2.com/research/sota/medical-question-answering-on-genmedgpt-5k-1-0-test.md?metric=Disease%20Diagnosis%20Loss%20Rate>): Lower is better; unit: PERCENT.

- [Drug Recommendation Win Rate](<https://www.sota2.com/research/sota/medical-question-answering-on-genmedgpt-5k-1-0-test.md?metric=Drug%20Recommendation%20Win%20Rate>): Higher is better; unit: PERCENT.

- [Drug Recommendation Tie Rate](<https://www.sota2.com/research/sota/medical-question-answering-on-genmedgpt-5k-1-0-test.md?metric=Drug%20Recommendation%20Tie%20Rate>): Lower is better; unit: PERCENT.

- [Drug Recommendation Loss Rate](<https://www.sota2.com/research/sota/medical-question-answering-on-genmedgpt-5k-1-0-test.md?metric=Drug%20Recommendation%20Loss%20Rate>): Lower is better; unit: PERCENT.

- [Overall Win Rate](<https://www.sota2.com/research/sota/medical-question-answering-on-genmedgpt-5k-1-0-test.md?metric=Overall%20Win%20Rate>): Higher is better; unit: PERCENT.

- [Overall Tie Rate](<https://www.sota2.com/research/sota/medical-question-answering-on-genmedgpt-5k-1-0-test.md?metric=Overall%20Tie%20Rate>): Lower is better; unit: PERCENT.

- [Overall Loss Rate](<https://www.sota2.com/research/sota/medical-question-answering-on-genmedgpt-5k-1-0-test.md?metric=Overall%20Loss%20Rate>): Lower is better; unit: PERCENT.

## Recorded results

| Method | Diversity & Integrity Win Rate | Result date | Report paper |
| --- | --- | --- | --- |
| MindMap | 100 | Not recorded | 68fa72631e209131cea56268 |
| MindMap | 100 | Not recorded | 68fa72631e209131cea56268 |
| MindMap | 100 | Not recorded | 68fa72631e209131cea56268 |
| MindMap | 100 | Not recorded | 68fa72631e209131cea56268 |
| MindMap | 100 | Not recorded | 68fa72631e209131cea56268 |
| MindMap | 100 | Not recorded | 68fa72631e209131cea56268 |

### MindMap

Diversity & Integrity Win Rate: 100. Result ID: 699236423a6c3148490d57ec.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72631e209131cea56268.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72631e209131cea56268.md>)

[Implementation](<https://github.com/wyl-willing/MindMap>)

Recorded evaluation configuration: {"Baseline Model (Comparator)":"GPT-3.5","Evaluation Rater":"GPT-4"}. This record does not assert independent reproduction.

### MindMap

Diversity & Integrity Win Rate: 100. Result ID: 699236423a6c3148490d57ed.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72631e209131cea56268.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72631e209131cea56268.md>)

[Implementation](<https://github.com/wyl-willing/MindMap>)

Recorded evaluation configuration: {"Baseline Model (Comparator)":"BM25 Retriever","Evaluation Rater":"GPT-4"}. This record does not assert independent reproduction.

### MindMap

Diversity & Integrity Win Rate: 100. Result ID: 699236423a6c3148490d57ee.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72631e209131cea56268.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72631e209131cea56268.md>)

[Implementation](<https://github.com/wyl-willing/MindMap>)

Recorded evaluation configuration: {"Baseline Model (Comparator)":"Embedding Retriever","Evaluation Rater":"GPT-4"}. This record does not assert independent reproduction.

### MindMap

Diversity & Integrity Win Rate: 100. Result ID: 699236423a6c3148490d57ef.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72631e209131cea56268.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72631e209131cea56268.md>)

[Implementation](<https://github.com/wyl-willing/MindMap>)

Recorded evaluation configuration: {"Baseline Model (Comparator)":"KG Retriever","Evaluation Rater":"GPT-4"}. This record does not assert independent reproduction.

### MindMap

Diversity & Integrity Win Rate: 100. Result ID: 699236423a6c3148490d57f0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72631e209131cea56268.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72631e209131cea56268.md>)

[Implementation](<https://github.com/wyl-willing/MindMap>)

Recorded evaluation configuration: {"Baseline Model (Comparator)":"GPT-4","Evaluation Rater":"GPT-4"}. This record does not assert independent reproduction.

### MindMap

Diversity & Integrity Win Rate: 100. Result ID: 699236423a6c3148490d57f1.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72631e209131cea56268.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72631e209131cea56268.md>)

[Implementation](<https://github.com/wyl-willing/MindMap>)

Recorded evaluation configuration: {"Baseline Model (Comparator)":"TOT","Evaluation Rater":"GPT-4"}. This record does not assert independent reproduction.

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