# Open-Ended Question Answering on Earth Observation

Source: [Open-Ended Question Answering on Earth Observation](<https://www.sota2.com/research/sota/open-ended-question-answering-on-earth-observation>)

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

Data updated: 2026-04-16T06:34:12.525Z

Evaluation of model performance on the Earth Observation dataset for Open-Ended Question Answering, measured using Judge scores, EVE WR, and Rank.

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 Question Answering |
| Dataset (including any stated split) | Earth Observation |
| Metric | Judge Score |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 7 of 7 |
| 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

- [Judge Score](<https://www.sota2.com/research/sota/open-ended-question-answering-on-earth-observation.md?metric=Judge%20Score>): Higher is better; unit: PERCENT.

- [EVE WR](<https://www.sota2.com/research/sota/open-ended-question-answering-on-earth-observation.md?metric=EVE%20WR>): Higher is better; unit: PERCENT.

- [Rank](<https://www.sota2.com/research/sota/open-ended-question-answering-on-earth-observation.md?metric=Rank>): Lower is better; unit: SCALAR.

## Recorded results

| Method | Judge Score | Result date | Report paper |
| --- | --- | --- | --- |
| Qwen3 | 97.05 | Not recorded | 69e06ce27d0172140ec7c876 |
| GPT-4.1 | 96.48 | Not recorded | 69e06ce27d0172140ec7c876 |
| Mistral Medium 3.1 | 96.45 | Not recorded | 69e06ce27d0172140ec7c876 |
| EVE-Instruct | 96.4 | Not recorded | 69e06ce27d0172140ec7c876 |
| GPT OSS | 94.2 | Not recorded | 69e06ce27d0172140ec7c876 |
| GPT-5 nano | 92.2 | Not recorded | 69e06ce27d0172140ec7c876 |
| MiniMax m2.5 | 91 | Not recorded | 69e06ce27d0172140ec7c876 |

### Qwen3

Judge Score: 97.05. Result ID: 69e082cb47f34ce924c91f4d.

[Reporting paper](<https://www.sota2.com/research/paper/69e06ce27d0172140ec7c876.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79601e209131cea7d516.md>)

Recorded evaluation configuration: {"Size (B)":"235-A22"}. This record does not assert independent reproduction.

### GPT-4.1

Judge Score: 96.48. Result ID: 69e082cb47f34ce924c91f4c.

[Reporting paper](<https://www.sota2.com/research/paper/69e06ce27d0172140ec7c876.md>)

Recorded evaluation configuration: {"Size (B)":"1800\*"}. This record does not assert independent reproduction.

### Mistral Medium 3.1

Judge Score: 96.45. Result ID: 69e082cb47f34ce924c91f50.

[Reporting paper](<https://www.sota2.com/research/paper/69e06ce27d0172140ec7c876.md>)

Recorded evaluation configuration: {"Size (B)":"200\*"}. This record does not assert independent reproduction.

### EVE-Instruct

Judge Score: 96.4. Result ID: 69e082cb47f34ce924c91f52.

[Reporting paper](<https://www.sota2.com/research/paper/69e06ce27d0172140ec7c876.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69e06ce27d0172140ec7c876.md>)

Recorded evaluation configuration: {"Size (B)":"24"}. This record does not assert independent reproduction.

### GPT OSS

Judge Score: 94.2. Result ID: 69e082cb47f34ce924c91f4f.

[Reporting paper](<https://www.sota2.com/research/paper/69e06ce27d0172140ec7c876.md>)

Recorded evaluation configuration: {"Size (B)":"120A5"}. This record does not assert independent reproduction.

### GPT-5 nano

Judge Score: 92.2. Result ID: 69e082cb47f34ce924c91f51.

[Reporting paper](<https://www.sota2.com/research/paper/69e06ce27d0172140ec7c876.md>)

Recorded evaluation configuration: {"Size (B)":"20\*"}. This record does not assert independent reproduction.

### MiniMax m2.5

Judge Score: 91. Result ID: 69e082cb47f34ce924c91f4e.

[Reporting paper](<https://www.sota2.com/research/paper/69e06ce27d0172140ec7c876.md>)

Recorded evaluation configuration: {"Size (B)":"230A10"}. This record does not assert independent reproduction.

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