# GIS agentic spatial analysis on GeoAgentBench (GABench)

Source: [GIS agentic spatial analysis on GeoAgentBench (GABench)](<https://www.sota2.com/research/sota/gis-agentic-spatial-analysis-on-geoagentbench-gabench>)

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

Data updated: 2026-04-16T06:12:58.317Z

Evaluation of agent performance on GIS agentic spatial analysis tasks using the GeoAgentBench (GABench) dataset, assessed via metrics like TAO scores, TIO, TEM, and efficiency measures.

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 | GIS agentic spatial analysis |
| Dataset (including any stated split) | GeoAgentBench (GABench) |
| Metric | TAO (Reasoning) |
| 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

- [TAO (Reasoning)](<https://www.sota2.com/research/sota/gis-agentic-spatial-analysis-on-geoagentbench-gabench.md?metric=TAO%20(Reasoning)>): Higher is better; unit: PERCENT.

- [TAO (Precision)](<https://www.sota2.com/research/sota/gis-agentic-spatial-analysis-on-geoagentbench-gabench.md?metric=TAO%20(Precision)>): Higher is better; unit: PERCENT.

- [TAO (F1 Score)](<https://www.sota2.com/research/sota/gis-agentic-spatial-analysis-on-geoagentbench-gabench.md?metric=TAO%20(F1%20Score)>): Higher is better; unit: PERCENT.

- [TIO](<https://www.sota2.com/research/sota/gis-agentic-spatial-analysis-on-geoagentbench-gabench.md?metric=TIO>): Higher is better; unit: PERCENT.

- [TEM](<https://www.sota2.com/research/sota/gis-agentic-spatial-analysis-on-geoagentbench-gabench.md?metric=TEM>): Higher is better; unit: PERCENT.

- [PEA Score](<https://www.sota2.com/research/sota/gis-agentic-spatial-analysis-on-geoagentbench-gabench.md?metric=PEA%20Score>): Higher is better; unit: PERCENT.

- [VLM Score](<https://www.sota2.com/research/sota/gis-agentic-spatial-analysis-on-geoagentbench-gabench.md?metric=VLM%20Score>): Higher is better; unit: PERCENT.

- [Efficiency (Macro)](<https://www.sota2.com/research/sota/gis-agentic-spatial-analysis-on-geoagentbench-gabench.md?metric=Efficiency%20(Macro)>): Higher is better; unit: PERCENT.

- [Efficiency (Micro)](<https://www.sota2.com/research/sota/gis-agentic-spatial-analysis-on-geoagentbench-gabench.md?metric=Efficiency%20(Micro)>): Higher is better; unit: PERCENT.

## Recorded results

| Method | TAO (Reasoning) | Result date | Report paper |
| --- | --- | --- | --- |
| Claude 3.5 | 89 | Not recorded | 69e06d077d0172140ec7ca3d |
| DeepSeek-V3 | 83.79 | Not recorded | 69e06d077d0172140ec7ca3d |
| GPT-4o | 79.4 | Not recorded | 69e06d077d0172140ec7ca3d |
| Gemini-1.5 | 75.99 | Not recorded | 69e06d077d0172140ec7ca3d |
| GPT-4o-mini | 74.08 | Not recorded | 69e06d077d0172140ec7ca3d |
| Llama-3.1-8B | 44.26 | Not recorded | 69e06d077d0172140ec7ca3d |
| Qwen2.5-7B | 41.86 | Not recorded | 69e06d077d0172140ec7ca3d |

### Claude 3.5

TAO (Reasoning): 89. Result ID: 69e07da147f34ce924c9126a.

[Reporting paper](<https://www.sota2.com/research/paper/69e06d077d0172140ec7ca3d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69e06d077d0172140ec7ca3d.md>)

Recorded evaluation configuration: {"Framework":"Plan-and-React"}. This record does not assert independent reproduction.

### DeepSeek-V3

TAO (Reasoning): 83.79. Result ID: 69e07da147f34ce924c91269.

[Reporting paper](<https://www.sota2.com/research/paper/69e06d077d0172140ec7ca3d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69e06d077d0172140ec7ca3d.md>)

Recorded evaluation configuration: {"Framework":"Plan-and-React"}. This record does not assert independent reproduction.

### GPT-4o

TAO (Reasoning): 79.4. Result ID: 69e07da147f34ce924c91268.

[Reporting paper](<https://www.sota2.com/research/paper/69e06d077d0172140ec7ca3d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69e06d077d0172140ec7ca3d.md>)

Recorded evaluation configuration: {"Framework":"Plan-and-React"}. This record does not assert independent reproduction.

### Gemini-1.5

TAO (Reasoning): 75.99. Result ID: 69e07da147f34ce924c9126b.

[Reporting paper](<https://www.sota2.com/research/paper/69e06d077d0172140ec7ca3d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69e06d077d0172140ec7ca3d.md>)

Recorded evaluation configuration: {"Framework":"Plan-and-React"}. This record does not assert independent reproduction.

### GPT-4o-mini

TAO (Reasoning): 74.08. Result ID: 69e07da147f34ce924c91267.

[Reporting paper](<https://www.sota2.com/research/paper/69e06d077d0172140ec7ca3d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69e06d077d0172140ec7ca3d.md>)

Recorded evaluation configuration: {"Framework":"Plan-and-React"}. This record does not assert independent reproduction.

### Llama-3.1-8B

TAO (Reasoning): 44.26. Result ID: 69e07da147f34ce924c91266.

[Reporting paper](<https://www.sota2.com/research/paper/69e06d077d0172140ec7ca3d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69e06d077d0172140ec7ca3d.md>)

Recorded evaluation configuration: {"Framework":"Plan-and-React"}. This record does not assert independent reproduction.

### Qwen2.5-7B

TAO (Reasoning): 41.86. Result ID: 69e07da147f34ce924c91265.

[Reporting paper](<https://www.sota2.com/research/paper/69e06d077d0172140ec7ca3d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69e06d077d0172140ec7ca3d.md>)

Recorded evaluation configuration: {"Framework":"Plan-and-React"}. This record does not assert independent reproduction.

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