Video-MemComp OVO Allocation Weights - Forward Average: leaderboard
Video-MemComp v1: streaming evaluation of official development sets. Table 1 mixes reproduced starred rows and results from prior papers; frame sampling and average KV budget differ by system and remain in row identity. Backbone checkpoint revision and decoding are not fully specified; source-qualified systems are retained. This is video QA under cache compression, not an unmodified backbone score. Appendix F identifies Qwen2.5-VL-7B on OVO. These are distinct hyperparameter arms evaluated on the same questions, not independently sampled tasks; preserve every configuration. Forward Average. Metric: Accuracy (%).
Metric: Accuracy (%). Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 27 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Video-MemComp Qwen2.5-VL-7B Allocation Weights [0.275, 0.675, 0.05] | 41.46 | |
| 2 | Video-MemComp Qwen2.5-VL-7B Allocation Weights [0.30, 0.675, 0.025] | 41.45 | |
| 3 | Video-MemComp Qwen2.5-VL-7B Allocation Weights [0.30, 0.65, 0.05] | 41.22 | |
| 4 | Video-MemComp Qwen2.5-VL-7B Allocation Weights [0.70, 0.20, 0.10] | 41.17 | |
| 5 | Video-MemComp Qwen2.5-VL-7B Allocation Weights [0.35, 0.625, 0.025] | 41.16 | |
| 6 | Video-MemComp Qwen2.5-VL-7B Allocation Weights [0.50, 0.45, 0.05] | 40.92 | |
| 7 | Video-MemComp Qwen2.5-VL-7B Allocation Weights [0.50, 0.40, 0.10] | 40.88 | |
| 8 | Video-MemComp Qwen2.5-VL-7B Allocation Weights [0.10, 0.40, 0.50] | 40.72 | |
| 9 | Video-MemComp Qwen2.5-VL-7B Allocation Weights [0.30, 0.30, 0.40] | 40.68 | |
| 10 | Video-MemComp Qwen2.5-VL-7B Allocation Weights [0.00, 0.00, 1.00] | 40.67 |
Interactive version: theaggregate.ai/benchmark?slug=video-memcomp-ovo-allocation-weights-forward-average · How It Works · Data refreshed daily, snapshot 2026-10-27.