Metric Depth Estimation on ETH3D (delta1)
90.8Delta-1 AccuracyMoGe-2
Evaluation Results
| Method | Links | |
|---|---|---|
| MoGe-22026.05 | 90.8 | |
| Metric3Dv22026.05 | 90 | |
| UnidepthV22026.05 | 85.2 | |
| Ours-4bParameters=4b2026.05 | 81 | |
| VLM3-4bModel Type=VLM Trained on Metric Depth Estimation2026.05 | 81 | |
| DepthLM-3bModel Type=VLM Trained on Metric Depth Estimation2026.05 | 74.5 | |
| DepthLM-7bModel Type=VLM Trained on Metric Depth Estimation2026.05 | 71.8 | |
| Metric3D2026.05 | 45.6 | |
| Depth Pro2026.05 | 39.7 | |
| DepthAnythingV22026.05 | 36.3 | |
| ZoeDepth2026.05 | 35 | |
| Gemini-2.5-ProModel Type=VLM2026.05 | 32.8 | |
| GPT-5Model Type=VLM2026.05 | 31.3 | |
| Seed1.5-VLModel Type=VLM Trained on Metric Depth Estimation2026.05 | 30.9 | |
| Qwen2.5-VL-72bModel Type=VLM2026.05 | 22 | |
| SpatialRGPT-8BModel Type=Spatial VLM2026.05 | 22 | |
| Unidepth2026.05 | 18.5 | |
| Qwen3-VL-32bModel Type=VLM2026.05 | 16.7 | |
| DepthAnything2026.05 | 9.3 | |
| SpaceLLaVA-13BModel Type=Spatial VLM2026.05 | 9 | |
| Qwen3-VL-4bModel Type=VLM2026.05 | 7.1 |