Vision-and-Language Navigation on VLN-CE (test-unseen)
5.15Navigation Error (NE)ETPNav
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| ETPNavTraining paradigm=Supervised2026.01 | 5.15 | — | 58 | 52 | 53.18 | 61.15 | |
| RecBERTTraining paradigm=Supervised2026.01 | 5.8 | — | 57 | 48 | 43.22 | 54.81 | |
| CWP-VLNBERTPanoramic images=true2022.10 | 5.89 | 13.31 | 51 | 42 | 36 | — | |
| CWP-CMAPanoramic images=true2022.10 | 6.3 | 11.85 | 49 | 38 | 33 | — | |
| WaypointTeamPanoramic images=true2022.10 | 6.65 | 8.02 | 37 | 32 | 30 | — | |
| Spatial-VLN (Deepseekv3)Training paradigm=Zero-Shot, Backbone=Deepseekv32026.01 | 6.65 | — | 34 | 33 | 27.44 | 47.37 | |
| Open-Nav-GPT4Training paradigm=Zero-Shot2026.01 | 6.7 | — | 23 | 19 | 16.1 | 45.79 | |
| SmartWay-GPT4oTraining paradigm=Zero-Shot2026.01 | 7.01 | — | 51 | 29 | 22.46 | — | |
| WS-MGMapPanoramic images=false2022.10 | 7.11 | 12.3 | 45 | 35 | 28 | — | |
| Open-Nav-Llama3.1Training paradigm=Zero-Shot2026.01 | 7.25 | — | 23 | 16 | 12.9 | 44.99 | |
| Spatial-VLN (GPT-OSS-20B)Training paradigm=Zero-Shot, Backbone=GPT-OSS-20B2026.01 | 7.33 | — | 32 | 25 | 21.5 | 46.95 | |
| CA-Nav-GPT4Training paradigm=Zero-Shot2026.01 | 7.58 | — | 48 | 25.3 | 10.8 | — | |
| CM2Panoramic images=false2022.10 | 7.74 | 13.85 | 39 | 31 | 24 | — | |
| DisscussNav-GPT4Training paradigm=Zero-Shot2026.01 | 7.77 | — | 15 | 11 | 10.51 | 42.87 | |
| VIRL_TeamPanoramic images=false2022.10 | 7.91 | 8.85 | 36 | 28 | 25 | — | |
| RandomTraining paradigm=Zero-Shot2026.01 | 8.63 | — | 12 | 2 | 1.5 | 34.08 | |
| InstructNav-GPT4Training paradigm=Zero-Shot2026.01 | 9.2 | — | 47 | 17 | 11 | — |