# Aerial Visual-Spatial Reasoning on UAV aerial-view benchmark

Source: [Aerial Visual-Spatial Reasoning on UAV aerial-view benchmark](<https://www.sota2.com/research/sota/aerial-visual-spatial-reasoning-on-uav-aerial-view-benchmark>)

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

Data updated: 2026-06-11T10:46:41.851Z

Evaluation of model performance on the UAV aerial-view benchmark dataset for Aerial Visual-Spatial Reasoning tasks, reporting scores across Overall, Single, Multi, Binary, Interval, and Ranking metrics.

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 | Aerial Visual-Spatial Reasoning |
| Dataset (including any stated split) | UAV aerial-view benchmark |
| Metric | Overall Score |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 14 of 14 |
| 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

- [Overall Score](<https://www.sota2.com/research/sota/aerial-visual-spatial-reasoning-on-uav-aerial-view-benchmark.md?metric=Overall%20Score>): Higher is better; unit: PERCENT.

- [Single Score](<https://www.sota2.com/research/sota/aerial-visual-spatial-reasoning-on-uav-aerial-view-benchmark.md?metric=Single%20Score>): Higher is better; unit: PERCENT.

- [Multi Score](<https://www.sota2.com/research/sota/aerial-visual-spatial-reasoning-on-uav-aerial-view-benchmark.md?metric=Multi%20Score>): Higher is better; unit: PERCENT.

- [Binary Score](<https://www.sota2.com/research/sota/aerial-visual-spatial-reasoning-on-uav-aerial-view-benchmark.md?metric=Binary%20Score>): Higher is better; unit: PERCENT.

- [Interval Score](<https://www.sota2.com/research/sota/aerial-visual-spatial-reasoning-on-uav-aerial-view-benchmark.md?metric=Interval%20Score>): Higher is better; unit: PERCENT.

- [Ranking Score](<https://www.sota2.com/research/sota/aerial-visual-spatial-reasoning-on-uav-aerial-view-benchmark.md?metric=Ranking%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Overall Score | Result date | Report paper |
| --- | --- | --- | --- |
| GPT-5.5 | 74.13 | Not recorded | 6a2a4109d33dfb23e783929f |
| Gemini-3-Pro-Preview | 71.83 | Not recorded | 6a2a4109d33dfb23e783929f |
| Kimi-K2.5 | 71.8 | Not recorded | 6a2a4109d33dfb23e783929f |
| Qwen3.6-27B | 71.27 | Not recorded | 6a2a4109d33dfb23e783929f |
| Qwen3.6-35B-A3B | 69 | Not recorded | 6a2a4109d33dfb23e783929f |
| Grok-4.20-0309-Reasoning | 68.57 | Not recorded | 6a2a4109d33dfb23e783929f |
| Moonshot-VL-128K-Vision-Preview | 67.53 | Not recorded | 6a2a4109d33dfb23e783929f |
| Claude-Opus-4-7 | 67.23 | Not recorded | 6a2a4109d33dfb23e783929f |
| Claude-Sonnet-4-6 | 67.07 | Not recorded | 6a2a4109d33dfb23e783929f |
| GPT-5.2 | 65.87 | Not recorded | 6a2a4109d33dfb23e783929f |
| GPT-4.1 | 63.13 | Not recorded | 6a2a4109d33dfb23e783929f |
| Qwen3-VL-Plus | 56.87 | Not recorded | 6a2a4109d33dfb23e783929f |
| Claude-Haiku-4-5-20251001 | 51.33 | Not recorded | 6a2a4109d33dfb23e783929f |
| Qwen3-VL-Flash-2025-10-15 | 49.87 | Not recorded | 6a2a4109d33dfb23e783929f |

### GPT-5.5

Overall Score: 74.13. Result ID: 6a2a91d9fbc373592c288ac2.

[Reporting paper](<https://www.sota2.com/research/paper/6a2a4109d33dfb23e783929f.md>)

Recorded evaluation configuration: {"Source Type":"Closed-source / API","Setting":"vision-language"}. This record does not assert independent reproduction.

### Gemini-3-Pro-Preview

Overall Score: 71.83. Result ID: 6a2a91d9fbc373592c288ac3.

[Reporting paper](<https://www.sota2.com/research/paper/6a2a4109d33dfb23e783929f.md>)

Recorded evaluation configuration: {"Source Type":"Closed-source / API","Setting":"vision-language"}. This record does not assert independent reproduction.

### Kimi-K2.5

Overall Score: 71.8. Result ID: 6a2a91d9fbc373592c288ac4.

[Reporting paper](<https://www.sota2.com/research/paper/6a2a4109d33dfb23e783929f.md>)

Recorded evaluation configuration: {"Source Type":"Closed-source / API","Setting":"vision-language"}. This record does not assert independent reproduction.

### Qwen3.6-27B

Overall Score: 71.27. Result ID: 6a2a91d9fbc373592c288ace.

[Reporting paper](<https://www.sota2.com/research/paper/6a2a4109d33dfb23e783929f.md>)

Recorded evaluation configuration: {"Source Type":"Open-source / Local","Setting":"vision-language"}. This record does not assert independent reproduction.

### Qwen3.6-35B-A3B

Overall Score: 69. Result ID: 6a2a91d9fbc373592c288acf.

[Reporting paper](<https://www.sota2.com/research/paper/6a2a4109d33dfb23e783929f.md>)

Recorded evaluation configuration: {"Source Type":"Open-source / Local","Setting":"vision-language"}. This record does not assert independent reproduction.

### Grok-4.20-0309-Reasoning

Overall Score: 68.57. Result ID: 6a2a91d9fbc373592c288ac5.

[Reporting paper](<https://www.sota2.com/research/paper/6a2a4109d33dfb23e783929f.md>)

Recorded evaluation configuration: {"Source Type":"Closed-source / API","Setting":"vision-language"}. This record does not assert independent reproduction.

### Moonshot-VL-128K-Vision-Preview

Overall Score: 67.53. Result ID: 6a2a91d9fbc373592c288ac6.

[Reporting paper](<https://www.sota2.com/research/paper/6a2a4109d33dfb23e783929f.md>)

Recorded evaluation configuration: {"Source Type":"Closed-source / API","Setting":"vision-language"}. This record does not assert independent reproduction.

### Claude-Opus-4-7

Overall Score: 67.23. Result ID: 6a2a91d9fbc373592c288ac7.

[Reporting paper](<https://www.sota2.com/research/paper/6a2a4109d33dfb23e783929f.md>)

Recorded evaluation configuration: {"Source Type":"Closed-source / API","Setting":"vision-language"}. This record does not assert independent reproduction.

### Claude-Sonnet-4-6

Overall Score: 67.07. Result ID: 6a2a91d9fbc373592c288ac8.

[Reporting paper](<https://www.sota2.com/research/paper/6a2a4109d33dfb23e783929f.md>)

Recorded evaluation configuration: {"Source Type":"Closed-source / API","Setting":"vision-language"}. This record does not assert independent reproduction.

### GPT-5.2

Overall Score: 65.87. Result ID: 6a2a91d9fbc373592c288ac9.

[Reporting paper](<https://www.sota2.com/research/paper/6a2a4109d33dfb23e783929f.md>)

Recorded evaluation configuration: {"Source Type":"Closed-source / API","Setting":"vision-language"}. This record does not assert independent reproduction.

### GPT-4.1

Overall Score: 63.13. Result ID: 6a2a91d9fbc373592c288aca.

[Reporting paper](<https://www.sota2.com/research/paper/6a2a4109d33dfb23e783929f.md>)

Recorded evaluation configuration: {"Source Type":"Closed-source / API","Setting":"vision-language"}. This record does not assert independent reproduction.

### Qwen3-VL-Plus

Overall Score: 56.87. Result ID: 6a2a91d9fbc373592c288acb.

[Reporting paper](<https://www.sota2.com/research/paper/6a2a4109d33dfb23e783929f.md>)

Recorded evaluation configuration: {"Source Type":"Closed-source / API","Setting":"vision-language"}. This record does not assert independent reproduction.

### Claude-Haiku-4-5-20251001

Overall Score: 51.33. Result ID: 6a2a91d9fbc373592c288acc.

[Reporting paper](<https://www.sota2.com/research/paper/6a2a4109d33dfb23e783929f.md>)

Recorded evaluation configuration: {"Source Type":"Closed-source / API","Setting":"vision-language"}. This record does not assert independent reproduction.

### Qwen3-VL-Flash-2025-10-15

Overall Score: 49.87. Result ID: 6a2a91d9fbc373592c288acd.

[Reporting paper](<https://www.sota2.com/research/paper/6a2a4109d33dfb23e783929f.md>)

Recorded evaluation configuration: {"Source Type":"Closed-source / API","Setting":"vision-language"}. This record does not assert independent reproduction.

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