AdaCodec: A Predictive Visual Code for Video MLLMs
This work addresses the token redundancy problem in video MLLMs, offering a more efficient video interface that reduces computational cost without sacrificing performance.
AdaCodec introduces a predictive visual code for video MLLMs that encodes inter-frame changes instead of treating each frame as an independent RGB image, reducing visual token redundancy. At 1/7 the token budget, it surpasses the 224k baseline on all long-video benchmarks and cuts time-to-first-token from 9.26s to 1.62s while improving average scores on general-video benchmarks.
Video is temporally redundant: adjacent frames usually share most objects, background, and layout. Yet existing video multimodal large language models (video MLLMs) usually encode each sampled frame as an independent RGB image, causing visual tokens to repeat content already present in earlier frames. This suggests a more direct video interface: send a full reference frame only when the scene cannot be predicted well from prior context, and otherwise transmit a compact description of inter-frame changes. We call this interface a \emph{predictive visual code}, and instantiate it for video MLLMs as \textbf{AdaCodec}. AdaCodec spends full visual tokens on a reference frame only when its conditional predictive cost is high; otherwise, it encodes inter-frame changes, including motion and prediction residuals, as compact P-tokens. Across all eleven benchmarks, AdaCodec improves over the Qwen3-VL-8B per-frame RGB baseline at a matched visual-token budget. Even at $1/7$ the budget, AdaCodec with 32k tokens surpasses the 224k baseline on all long-video benchmarks; on five general-video benchmarks, it raises the average score while substantially cutting time-to-first-token from 9.26s to 1.62s.