Polygon Area Estimation on SpatialEval Beijing 1.0 (test)
1,050.9MAESRL
Evaluation results
| Method | Links | ||
|---|---|---|---|
| SRLdescription=Spatial Representation Learning upper bound2026.06 | 1,050.9 | — | |
| gemini-2.5-flash2026.06 | 1,217.3 | 100 | |
| ChatGPT 5.12026.06 | 1,314.9 | 100 | |
| ChatGPT 5.1mode=In-Context2026.06 | 1,314.9 | 100 | |
| SLMbase_model=Qwen3-8B, fine-tuning=LoRA2026.06 | 1,376 | 100 | |
| Qwen3-8Bmode=tuned2026.06 | 1,809.8 | 100 | |
| SpatialRGPT2026.06 | 2,092.62 | 100 | |
| Qwen3-8B2026.06 | 4,093.5 | 100 | |
| DeepSeek-R1-32B2026.06 | 5,293.7 | 13.3 | |
| Llama-3.3-70B2026.06 | 115,291.6 | 91.7 |