SleuthVLM: Do We Really Need Audio-in-Video for Omni Video Understanding?

peiran wu, Fangfei Gou, Zhaoxi Mu, Changhao Zhang, Jian Liu, Weiqiang Wang, Yuntao Wang, Yuanchun Shi, Junxiao Shen

SleuthVLM: Do We Really Need Audio-in-Video for Omni Video Understanding?: 6 upvotes on Hugging Face Daily Papers, #28 of 38 papers on 2026-11-05. Day-by-day upvote history.

Upvotes over time

It had 5 upvotes by its first day, 7 upvotes by the next day, 7 upvotes by day 3, 6 upvotes by day 7, and has 6 now. It ranked #143 of 207 papers in its week and #212 of 340 in its month. Its code repository has 2 GitHub stars.

Omni models answer questions about videos by jointly processing sampled frames and audio in one turn, with progress driven largely by scaling model size and omni training data. We ask whether a model needs to hear the audio directly at all, and present SleuthVLM, an agent harness built on a VLM that never hears the audio directly: speech will be transcribed into text, and the VLM will sleuth the video on its own without any predefined workflow. On four omni benchmarks with videos up to 90 minutes, SleuthVLM achieves the best average accuracy: 75.0%, 2.6 points above Gemini 3.8 Flash, and it also tops the Video-MME-v2 leaderboard with 77.8%; with an open-source 27B VLM, it beats all open-source omni models. We draw three findings: SleuthVLM beats a model that glances; hearing is not the bottleneck, agentic ability is; omni understanding needs better agentic ability, not just more omni data for modality alignment. So we provide a simple but effective starting point for doing omni video understanding. Omni understanding needs better agents, not just bigger models that hear.

Paper page on Hugging Face · arXiv

Data: hysts-bot-data/daily-papers-stats and the Daily Papers API. Open data: tardellirs/paper-pulse-data. Sister project: Model Pulse, the download history of every model on the Hub.