Video Question Answering on NExT-QA 5-180s (test)
80.54AccuracyInternVL3.5-8B + EC-RAG
Evaluation Results
| Method | Links | |
|---|---|---|
| InternVL3.5-8B + EC-RAGBackbone=InternVL3.5-8B, Augmentation=EC-RAG2026.05 | 80.54 | |
| VideoLlama3-7B + EC-RAGBackbone=VideoLlama3-7B, Augmentation=EC-RAG2026.05 | 78.48 | |
| InternVL3.5-8BBackbone=InternVL3.5-8B2026.05 | 77.54 | |
| GPT-4o (Closed-source)Access=Closed-source2026.05 | 76.71 | |
| VideoLlama3-7BBackbone=VideoLlama3-7B2026.05 | 76.69 | |
| Gemini1.5-pro (Closed-source)Access=Closed-source2026.05 | 75.74 | |
| Qwen3-VL-8B + EC-RAGBackbone=Qwen3-VL-8B, Augmentation=EC-RAG2026.05 | 75.54 | |
| Qwen3-VL-8BBackbone=Qwen3-VL-8B2026.05 | 73.08 |