Nearest Neighbor Query on SpatialEval Beijing
95.7PT-PL NN AccuracySRL
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| SRL2026.06 | 95.7 | 96.3 | — | — | |
| SLM2026.06 | 80.5 | 85.9 | 100 | 100 | |
| gemini-2.5-flash2026.06 | 71.9 | 75 | 100 | 100 | |
| ChatGPT 5.1In-Context Learning=true2026.06 | 67.2 | 70.3 | 100 | 100 | |
| ChatGPT 5.12026.06 | 66.4 | 68.8 | 100 | 100 | |
| Llama-3.3-70B2026.06 | 57 | 48.4 | 100 | 100 | |
| SpatialRGPT2026.06 | 33.6 | 28.1 | 100 | 100 | |
| GPT 5.1Entity Naming=true2026.06 | 30.46 | 25 | 88.3 | 93 | |
| gemini-2.5-flashEntity Naming=true2026.06 | 22.65 | 21.1 | 84.4 | 78.9 | |
| Qwen3-8BFine-tuning=true2026.06 | 20.31 | 28.91 | 100 | 100 | |
| Qwen3-8B2026.06 | 12.5 | 0.78 | 29.7 | 1.6 | |
| DeepSeek-R1-32B2026.06 | 10.2 | 60.9 | 10.9 | 63.3 |