AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

Video-Holmes: Can MLLM Think like Holmes for Complex Video Reasoning?

JUNHAO CHENG, Yuying Ge, Teng Wang, Yixiao Ge, Jing Liao, Ying Shan

ECCV 2026 · Vision-language and multimodal › Multimodal LLMs · approach: Chain-of-thought multimodal reasoning · 111 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 RL post-training for thinkingGemini-2.0-Flash · Video-Holmesaccuracy: 30.6 -> 43.1Improved
add RL post-training for thinkingGemini-2.0-Flash · MMVUaccuracy: 65.2 -> 64.5No clear effect
add RL post-training for thinkingGemini-2.0-Flash · Video-MMEaccuracy: 68.3 -> 69.4No clear effect
add RL video reasoning training (Video-R1)Qwen2.5-VL-7B · Video-Holmesoverall accuracy: 27.8 -> 36.5Improved
add RL training (SEED-Bench-R1)Qwen2.5-VL-7B · Video-Holmesoverall accuracy: 27.8 -> 33.5Improved
increase input frames from 32 to 80 (ablation)Qwen2.5-VL-7B · Video-Holmesaccuracy: 27.8 -> 33.0Improved
increase input frames from 32 to 50 (ablation)GPT-4o · Video-Holmesaccuracy: 42.0 -> 44.6Improved
add audio input (ablation)Gemini-2.5-Pro · Video-Holmesoverall accuracy: 45.0 -> 51.3Improved
add audio input (ablation)Qwen2.5-Omni-7B · Video-Holmesoverall accuracy: 16.4 -> 24.4Improved
enable CoT reasoning (ablation)Video-R1 · Video-Holmesaccuracy: 28.2 -> 36.5Improved
enable CoT reasoning (ablation)Qwen2.5-VL-7B · Video-Holmesaccuracy: 29.4 -> 27.8Hurt
enable CoT reasoning (ablation)Gemini-2.0-Flash · Video-Holmesaccuracy: 28.5 -> 30.6Improved

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