Labs
Every AI lab with a model in Geospatial Agent Benchmarks: its models, its best results and where they rank.
Every lab
| Lab | Open-weights models | Closed models | Best model, Geospatial Intelligence Index (Text only) | Score | Rank |
|---|---|---|---|---|---|
| Alibaba (Qwen) | 1 | 0 | Qwen 3.8 27B (xhigh) Open weights | 75.1 | #37 |
| Anthropic | 0 | 9 | Claude Opus 5.5 (xhigh) | 88.7 | #1 |
| DeepSeek | 3 | 0 | DeepSeek V4 Pro (max) Open weights | 79.0 | #23 |
| 1 | 3 | Gemini 3.8 Flash (low) | 78.7 | #25 | |
| Meta | 1 | 0 | Llama 3.3 70B (Non-reasoning) Open weights | 19.9 | #61 |
| MiniMax | 1 | 0 | MiniMax M3.1 Flash (max) Open weights | 60.5 | #49 |
| Mistral AI | 1 | 0 | Mistral Large 3 (Non-reasoning) Open weights | 44.5 | #55 |
| Moonshot AI | 3 | 0 | Kimi K3 (Reasoning) Open weights | 78.6 | #28 |
| NVIDIA | 4 | 0 | Nemotron 3 Ultra (high) Open weights | 73.7 | #39 |
| OpenAI | 2 | 6 | GPT-6.1 Sol (medium) | 86.4 | #3 |
| StepFun | 1 | 0 | Step 5 Preview (Reasoning) Open weights | 56.4 | #51 |
| Z.ai | 4 | 0 | GLM-5.2 (high) Open weights | 78.6 | #27 |
