Geocoding benchmark for geospatial AI
How well AI models do geocoding on satellite, aerial and map data: 2 sets with results, scored the same way for every model.
Sets
| Set | Index | Tasks of this type | Best score on the set | Models |
|---|---|---|---|---|
| Place and coordinate tasks, closed book | Spatial Intelligence Index | all | 92 | 22 |
| Place and coordinate tasks | Spatial Agent Index | all | 100 | 13 |
Top models
Mean of each model's best score on the 2 sets, 0 to 100, for the 13 models that have run all of them.
| Rank | Model | Developer | Score |
|---|---|---|---|
| 1 | Claude Opus 5.5 | Anthropic | 96 |
| 2 | Claude Sonnet 5.5 | Anthropic | 93 |
| 3 | Claude Haiku 5.5 | Anthropic | 85 |
| 4 | Nemotron 3 Ultra | NVIDIA | 84 |
| 5 | DeepSeek V4.1 Flash | DeepSeek | 80 |
| 6 | Step 5 Preview | StepFun | 80 |
| 7 | Muse Glimmer 30B | Meta | 76 |
| 8 | Mistral Large 4 | Mistral AI | 70 |
| 9 | MiMo V2.6 Flash | Xiaomi | 69 |
| 10 | MiniMax M3 | MiniMax | 67 |
Every task type · Every evaluation · Methodology
Frequently asked questions
Which AI model is best at geocoding on geospatial data?
Claude Opus 5.5 leads, on the sets below. Mean of each model's best score on the 2 sets, 0 to 100, for the 13 models that have run all of them.
What is this geocoding benchmark?
Every task in our sets that asks for geocoding, in GIS and Earth observation work, with each set's own score. It is a view over the sets, not an index of its own.
Which index does it count in?
Each set counts in one index: the Spatial Agent Index or the Spatial Intelligence Index, as the table shows.
