Video Question Answering and Temporal Grounding on ForeSeaQA-Text
67.7AccuracyVideoLLaMA3
Evaluation results
| Method | Links | ||
|---|---|---|---|
| VideoLLaMA3Params=7B2026.03 | 67.7 | 15.5 | |
| ForeSeaParams=7B2026.03 | 66.7 | 13.3 | |
| VideoRAGParams=7B2026.03 | 63.8 | 4.3 | |
| InternVL3Params=8B2026.03 | 63.4 | 9.9 | |
| InternVL3Params=9B2026.03 | 62.7 | 8.8 | |
| Qwen2.5-VLParams=72B2026.03 | 61.4 | 10.1 | |
| Qwen2.5-VLParams=7B2026.03 | 59 | 7.5 | |
| LLaVA-OneVisionParams=7B2026.03 | 58.5 | 7.7 | |
| GLM-4.1V-ThinkingParams=9B2026.03 | 55.1 | 8.4 | |
| T*Params=7B2026.03 | 48.4 | 4.2 | |
| InternVL3Params=2B2026.03 | 34.1 | 7.1 |