# Model Selection on HuggingGPT Human Evaluation Set 130 diverse requests (test)

Source: [Model Selection on HuggingGPT Human Evaluation Set 130 diverse requests (test)](<https://www.sota2.com/research/sota/model-selection-on-hugginggpt-human-evaluation-set-130-diverse-requests-test>)

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

Data updated: 2026-02-26T05:57:36.951Z

Evaluation of model selection performance on the HuggingGPT Human Evaluation Set 130 diverse requests (test split), measured by metrics like Passing Rate and Rationality.

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 | Model Selection |
| Dataset (including any stated split) | HuggingGPT Human Evaluation Set 130 diverse requests (test) |
| Metric | Passing Rate |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 1 of 1 |
| 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

- [Passing Rate](<https://www.sota2.com/research/sota/model-selection-on-hugginggpt-human-evaluation-set-130-diverse-requests-test.md?metric=Passing%20Rate>): Higher is better; unit: PERCENT.

- [Rationality](<https://www.sota2.com/research/sota/model-selection-on-hugginggpt-human-evaluation-set-130-diverse-requests-test.md?metric=Rationality>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Passing Rate | Result date | Report paper |
| --- | --- | --- | --- |
| HuggingGPT | 93.89 | Not recorded | 68fa71551e209131cea5091f |

### HuggingGPT

Passing Rate: 93.89. Result ID: 6991a7f43a6c31484909b2d0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71551e209131cea5091f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71551e209131cea5091f.md>)

[Implementation](<https://github.com/microsoft/JARVIS>)

Recorded evaluation configuration: {"LLM":"GPT-3.5"}. This record does not assert independent reproduction.

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