Multimodal Evaluation on MME (Total Score & Retention)
1,854.51Total ScoreShortV (N=20)
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ShortV (N=20)Model=LLaVA-NeXT, TFLOPs=51%2026.06 | 1,854.51 | 96 | |
| V2DropModel=LLaVA-NeXT, TFLOPs=70%2026.06 | 1,853.23 | 98.4 | |
| VTW (K=16)Model=LLaVA-NeXT, TFLOPs=51%2026.06 | 1,852.56 | 86.5 | |
| VanillaModel=LLaVA-NeXT, TFLOPs=100%2026.06 | 1,851.44 | 100 | |
| VanillaBase Model=LLaVA-NEXT-7B, Token Budget=2880, Pruning Rate=0%, Loc. (Pruning Location)=-2026.05 | 1,851 | 100 | |
| APETModel=LLaVA-NeXT, TFLOPs=70%2026.06 | 1,840.33 | 99.4 | |
| Ours (N=20)Model=LLaVA-NeXT, TFLOPs=67%2026.06 | 1,797.57 | 98.8 | |
| VSkip+ (N=20)Model=LLaVA-NeXT, TFLOPs=74%2026.06 | 1,759.67 | 98.2 | |
| HoloVBase Model=LLaVA-NEXT-7B, Token Budget=320, Pruning Rate=88.9%, Loc. (Pruning Location)=Pre2026.05 | 1,738 | 95.7 | |
| DiffPruneBase Model=LLaVA-NEXT-7B, Token Budget=320, Pruning Rate=88.9%, Loc. (Pruning Location)=Pre2026.05 | 1,723 | 96.1 | |
| DARTBase Model=LLaVA-NEXT-7B, Token Budget=320, Pruning Rate=88.9%, Loc. (Pruning Location)=L22026.05 | 1,710 | 95.6 | |
| HiREDBase Model=LLaVA-NEXT-7B, Token Budget=320, Pruning Rate=88.9%, Loc. (Pruning Location)=Pre2026.05 | 1,690 | 93.4 | |
| PDropBase Model=LLaVA-NEXT-7B, Token Budget=320, Pruning Rate=88.9%, Loc. (Pruning Location)=L8–L242026.05 | 1,663 | 90.7 | |
| FastVBase Model=LLaVA-NEXT-7B, Token Budget=320, Pruning Rate=88.9%, Loc. (Pruning Location)=L2–L322026.05 | 1,661 | 87.6 | |
| LLaVA-NeXT-7B (Upper Bound)Backbone=LLaVA-NeXT-7B, Tokens Kept=2880, Pruning Ratio=0.0%2025.12 | 1,513.8 | 100 | |
| FlashVLMBackbone=LLaVA-NeXT-7B, Tokens Kept=640, Pruning Ratio=77.8%2025.12 | 1,493.1 | 99.12 | |
| VisPrunerBackbone=LLaVA-NeXT-7B, Tokens Kept=640, Pruning Ratio=77.8%2025.12 | 1,480.7 | 98.6 | |
| FastVBackbone=LLaVA-NeXT-7B, Tokens Kept=640, Pruning Ratio=77.8%2025.12 | 1,477.3 | 97 | |
| VisionZipBackbone=LLaVA-NeXT-7B, Tokens Kept=640, Pruning Ratio=77.8%2025.12 | 1,468.4 | 96.7 | |
| SparseVLMBackbone=LLaVA-NeXT-7B, Tokens Kept=640, Pruning Ratio=77.8%2025.12 | 1,456.3 | 94.9 | |
| FlashVLMBackbone=LLaVA-NeXT-7B, Tokens Kept=320, Pruning Ratio=88.9%2025.12 | 1,396.7 | 94.61 | |
| VisPrunerBackbone=LLaVA-NeXT-7B, Tokens Kept=320, Pruning Ratio=88.9%2025.12 | 1,370.1 | 93.3 | |
| VisionZipBackbone=LLaVA-NeXT-7B, Tokens Kept=320, Pruning Ratio=88.9%2025.12 | 1,348.8 | 90.5 | |
| SparseVLMBackbone=LLaVA-NeXT-7B, Tokens Kept=320, Pruning Ratio=88.9%2025.12 | 1,342.7 | 87.9 | |
| FastVBackbone=LLaVA-NeXT-7B, Tokens Kept=320, Pruning Ratio=88.9%2025.12 | 1,282.9 | 87.7 | |
| FlashVLMBackbone=LLaVA-NeXT-7B, Tokens Kept=160, Pruning Ratio=94.4%2025.12 | 1,246.5 | 88.03 | |
| VisionZipBackbone=LLaVA-NeXT-7B, Tokens Kept=160, Pruning Ratio=94.4%2025.12 | 1,239.7 | 82.3 | |
| VisPrunerBackbone=LLaVA-NeXT-7B, Tokens Kept=160, Pruning Ratio=94.4%2025.12 | 1,226 | 86.7 | |
| SparseVLMBackbone=LLaVA-NeXT-7B, Tokens Kept=160, Pruning Ratio=94.4%2025.12 | 1,167.1 | 74.5 | |
| FastVBackbone=LLaVA-NeXT-7B, Tokens Kept=160, Pruning Ratio=94.4%2025.12 | 1,079.5 | 74.7 |