Video Question Answering on Video-MME (Length Breakdown)
65.3Accuracy (All)AKS
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| AKSBase MLLM=LLaVA-Video 7B, Sampled Frames=1fps, Input Tokens=13k2026.03 | 65.3 | 77.2 | 64.8 | 53.9 | |
| DToMABase MLLM=LLaVA-Video 7B, Input Tokens=12k2026.03 | 65 | — | — | — | |
| FlexMemBase MLLM=LLaVA-Video 7B, Sampled Frames=512/1024frm, Input Tokens=13k2026.03 | 64.7 | 76.3 | 63.3 | 54.4 | |
| PanelsBase MLLM=LLaVA-Video 7B, Sampled Frames=1fps, Input Tokens=13k2026.03 | 64.4 | — | 62.2 | 54 | |
| AdaRETAKEBase MLLM=LLaVA-Video 7B, Sampled Frames=1024frm, Input Tokens=40k2026.03 | 63.6 | 75.8 | 62 | 52.9 | |
| LLaVA-Video 7BBase MLLM=LLaVA-Video 7B, Sampled Frames=64frm, Input Tokens=13k2026.03 | 63.2 | 76.1 | 61 | 52.4 | |
| AdaRETAKEBase MLLM=LLaVA-OV 7B, Sampled Frames=1024frm, Input Tokens=20k2026.03 | 59 | 72.1 | 53.6 | 51.4 | |
| FlexMemBase MLLM=LLaVA-OV 7B, Sampled Frames=512/1024frm, Input Tokens=7k2026.03 | 59 | 70 | 57.3 | 49.8 | |
| PanelsBase MLLM=LLaVA-OV 7B, Sampled Frames=1fps, Input Tokens=7k2026.03 | 58.9 | — | 56.2 | 50.2 | |
| AKSBase MLLM=LLaVA-OV 7B, Sampled Frames=1fps, Input Tokens=7k2026.03 | 58.4 | — | — | — | |
| LLaVA-OV 7BBase MLLM=LLaVA-OV 7B, Sampled Frames=32frm, Input Tokens=7k2026.03 | 57.8 | 70.6 | 54.8 | 48.2 | |
| BOLTBase MLLM=LLaVA-OV 7B, Sampled Frames=1fps, Input Tokens=7k2026.03 | 57.8 | 69.2 | 56.8 | 47.3 |