Our new X account is live! Follow @wizwand_team for updates

Episodic-memory Question Answering on OpenEQA v1 (ScanNet)

87.7LLM-Match

Human baseline

Updated 4d ago

Evaluation Results

Method	Links
Human baseline 2025.12		87.7
Human baseline 2025.12		86.8
R4 2025.12		81.22
R4 2025.12		79.77
GPT-5 2025.12		64.4
GPT-5-mini 2025.12		59.2
3D-Mem 2025.12		57.2
GR-ER1.5 w/ Thinking 2025.12		55
GPT-4V 2025.12		51.3
GR-ER1.5 2025.12		50.5
GPT-4V 2025.12		49.6
AlanaVLM 2025.12		47.8
AlanaVLM 2025.12		46.7
GPT-4 w/ LLaVA-1.5 2025.12		45.4
GPT-4 w/ LLaVA-1.5 2025.12		43.6
GPT-4 w/ SVM 2025.12		40.9
LLaMA-2 w/ LLaVA-1.5 2025.12		39.6
GPT-4 w/ SVM 2025.12		38.9
GPT-4 w/ CG 2025.12		37.8
LLaMA-2 w/ LLaVA-1.5 2025.12		36.8
GPT-4 w/ CG 2025.12		36.5
LLaMA-2 w/ SVM 2025.12		36
LLaMA-2 w/ SVM 2025.12		34.3
GPT-4 2025.12		33.5
GPT-4 2025.12		32.5
LLaMA-2 w/ CG 2025.12		31
LLaMA-2 w/ CG 2025.12		28.7
LLaMA-2 2025.12		28.3
LLaMA-2 2025.12		27.9