AI Paper Atlas

Research map / Video / Video-language understanding

VideoChat-Flash: Hierarchical Compression for Long-Context Video Modeling

Xinhao Li, Yi Wang, Jiashuo Yu, Xiangyu Zeng, Yuhan Zhu, Haian Huang, Jianfei Gao, Kunchang Li, Yinan He, Chenting Wang, Yu Qiao, Yali Wang and 1 more

ICLR 2026 · Video › Video-language understanding · approach: Video LLMs with fine-grained understanding · 216 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add hierarchical compression HiCo (196 to 16 tokens/frame) (ablation)VideoChat-Flash baseline · MVBench / MLVUaccuracy: MVBench 60.2 -> 61.1; MLVU 63.7 -> 60.6Mixed
add timestamp prompt (ablation)VideoChat-Flash · LVBenchaccuracy: 44.2 -> 48.0Improved
add short video pretraining (ablation)VideoChat-Flash · MVBenchaccuracy: 61.1 -> 66.5Improved
add joint short and long video SFT (ablation)VideoChat-Flash · MLVUM-Avg: 68.3 -> 74.5Improved
use UMT-L video encoder instead of SigLIP at 16 tokens/frame (ablation)VideoChat-Flash · MVBench (8M data)accuracy: 71.2 -> 73.5Improved
use UMT-L instead of SigLIP with small data (2M) (ablation)VideoChat-Flash · MLVU (2M data)M-Avg: 62.0 -> 60.0Hurt
use similar token merging instead of spatial downsampling at 16 tokens/frame (ablation)connector · MVBench/MLVU/VideoMME avgavg accuracy: 56.4 -> 58.7Improved
use token merging at 196 tokens/frame vs uncompressed 729 tokens (ablation)connector · MVBench/MLVU/VideoMME avgavg accuracy: 59.6 -> 62.1Improved
add attention-based visual dropout at layer 18 (50% tokens) (ablation)Qwen2-7B video LLM · VideoMMEaccuracy: 61.2 -> 61.7Improved
drop 50% of visual tokens uniformly at shallow layer 4 (ablation)Qwen2-7B video LLM · VideoMMEaccuracy: 61.2 -> 60.4Hurt
swap UMT-L for InternVideo2-1B encoder (ablation)VideoChat-Flash · MLVUM-Avg: 74.5 -> 73.4Mixed

Later papers that compared with this method

6 later accepted papers report VideoChat-Flash as a baseline in their results tables; 5 of them report a better number on most of the results they share with it.

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy