Navigation Question Answering on OC-NaVQA
79.6Question AccuracyDAAAM + surprise EM
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| DAAAM + surprise EMReasoning Model=GPT-5-mini, Observation Time Tracking=true2026.06 | 79.6 | 36.57 | 1.51 | |
| DAAAMReasoning Model=GPT-5-mini, Observation Time Tracking=true2026.06 | 71.1 | 41.75 | 1.792 | |
| ReMEmbRBackbone=NVILA-Lite-8B, Reasoning Model=GPT-5-mini, Observation Time Tracking=true2026.06 | 46.3 | 55.894 | 4.106 | |
| ReMEmbRBackbone=NVILA-Lite-2B, Reasoning Model=GPT-5-mini, Observation Time Tracking=true2026.06 | 43.2 | 53.466 | 2.287 | |
| Concept-GraphsReasoning Model=GPT-5-mini, Observation Time Tracking=false2026.06 | 29.9 | 111.29 | — |