Scene-level Tasks on 3DCity-LLM 1.2M
20.11BLEU-43DCity-LLM
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| 3DCity-LLMTraining status=Training-required, Parameter scale=7B2026.03 | 20.11 | 29.87 | 37.84 | 7.5 | 6.16 | |
| City-VLMTraining status=Training-required, Parameter scale=7B2026.03 | 18.58 | 28.43 | 35.27 | 7.29 | 5.93 | |
| Chat SceneTraining status=Training-required, Parameter scale=7B2026.03 | 17.98 | 26.45 | 34.91 | 7.19 | 5.92 | |
| LLaVATraining status=Training-required, Parameter scale=7B2026.03 | 17.31 | 26.4 | 33.92 | 7.01 | 5.76 | |
| Chat 3DTraining status=Training-required, Parameter scale=7B2026.03 | 14.75 | 23.38 | 29.69 | 6.73 | 4.65 | |
| ChatGPT-5.1Training status=Training-free but payment-required, Parameter scale=>70B2026.03 | 11.44 | 24.26 | 35.63 | 7.27 | 6.04 | |
| Gemini 3Training status=Training-free but payment-required, Parameter scale=>70B2026.03 | 10.6 | 21.91 | 31.33 | 7.43 | 5.9 | |
| DeepSeek-R1Training status=Training-free but payment-required, Parameter scale=>70B2026.03 | 10.42 | 20.95 | 31.9 | 6.94 | 5.87 | |
| Qwen3-VL MaxTraining status=Training-free but payment-required, Parameter scale=>70B2026.03 | 9.32 | 21.63 | 32.37 | 7.4 | 6.12 |