# Temporal Chain of Thought: Long-Video Understanding by Thinking in Frames

arXiv: https://arxiv.org/abs/2507.02001 · HTML: https://mlsanity.com/paper/2507.02001/

## Key facts

- Cited by 27 papers in the map (data through 14 Oct 2026).
- Placed in Multimodal reasoning, one of 17 directions in LLM reasoning.
- Builds on 29 of its 74 references that are in the map.
- First posted on arXiv 1 Jul 2025.

## Placed in

- LLM reasoning / [Multimodal reasoning](https://mlsanity.com/lr/multimodal-reasoning/)
- Multimodal LLMs / vision-language models / [Video understanding with MLLMs](https://mlsanity.com/mllm/video-understanding-with-mllms/)

## Builds on (most cited first, 29 of 29)

- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning](https://mlsanity.com/paper/2501.12948/) (2025)
- [Self-Consistency Improves Chain of Thought Reasoning in Language Models](https://mlsanity.com/paper/2203.11171/) (2022)
- [Qwen2.5-VL Technical Report](https://mlsanity.com/paper/2502.13923/) (2025)
- [Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution](https://mlsanity.com/paper/2409.12191/) (2024)
- [Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context](https://mlsanity.com/paper/2403.05530/) (2024)
- [Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters](https://mlsanity.com/paper/2408.03314/) (2024)
- [Expanding Performance Boundaries of Open-Source Multimodal Models with Model, Data, and Test-Time Scaling](https://mlsanity.com/paper/2412.05271/) (2024)
- [RULER: What's the Real Context Size of Your Long-Context Language Models?](https://mlsanity.com/paper/2404.06654/) (2024)
- [Large Language Monkeys: Scaling Inference Compute with Repeated Sampling](https://mlsanity.com/paper/2407.21787/) (2024)
- [Habitat-Matterport 3D Dataset (HM3D): 1000 Large-scale 3D Environments for Embodied AI](https://mlsanity.com/paper/2109.08238/) (2021)
- [LLaVA-NeXT-Interleave: Tackling Multi-image, Video, and 3D in Large Multimodal Models](https://mlsanity.com/paper/2407.07895/) (2024)
- [LVBench: An Extreme Long Video Understanding Benchmark](https://mlsanity.com/paper/2406.08035/) (2024)
- [LongVU: Spatiotemporal Adaptive Compression for Long Video-Language Understanding](https://mlsanity.com/paper/2410.17434/) (2024)
- [VideoAgent: Long-form Video Understanding with Large Language Model as Agent](https://mlsanity.com/paper/2403.10517/) (2024)
- [Inference Scaling Laws: An Empirical Analysis of Compute-Optimal Inference for Problem-Solving with Language Models](https://mlsanity.com/paper/2408.00724/) (2024)
- [VideoTree: Adaptive Tree-based Video Representation for LLM Reasoning on Long Videos](https://mlsanity.com/paper/2405.19209/) (2024)
- [PIVOT: Iterative Visual Prompting Elicits Actionable Knowledge for VLMs](https://mlsanity.com/paper/2402.07872/) (2024)
- [Video-RAG: Visually-aligned Retrieval-Augmented Long Video Comprehension](https://mlsanity.com/paper/2411.13093/) (2024)
- [EVA-CLIP-18B: Scaling CLIP to 18 Billion Parameters](https://mlsanity.com/paper/2402.04252/) (2024)
- [LV-Eval: A Balanced Long-Context Benchmark with 5 Length Levels Up to 256K](https://mlsanity.com/paper/2402.05136/) (2024)
- [BOLT: Boost Large Vision-Language Model Without Training for Long-form Video Understanding](https://mlsanity.com/paper/2503.21483/) (2025)
- [Too Many Frames, Not All Useful: Efficient Strategies for Long-Form Video QA](https://mlsanity.com/paper/2406.09396/) (2024)
- [VideoEspresso: A Large-Scale Chain-of-Thought Dataset for Fine-Grained Video Reasoning via Core Frame Selection](https://mlsanity.com/paper/2411.14794/) (2024)
- [Structured information extraction from complex scientific text with fine-tuned large language models](https://mlsanity.com/paper/2212.05238/) (2022)
- [Language Repository for Long Video Understanding](https://mlsanity.com/paper/2403.14622/) (2024)
- [Visual Haystacks: A Vision-Centric Needle-In-A-Haystack Benchmark](https://mlsanity.com/paper/2407.13766/) (2024)
- [Needle In A Video Haystack: A Scalable Synthetic Evaluator for Video MLLMs](https://mlsanity.com/paper/2406.09367/) (2024)
- [Long Video Understanding with Learnable Retrieval in Video-Language Models](https://mlsanity.com/paper/2312.04931/) (2023)

## Built on by (most cited first, 27 of 27)

- [Active Video Perception: Iterative Evidence Seeking for Agentic Long Video Understanding](https://mlsanity.com/paper/2512.05774/) (2025)
- [ReWatch-R1: Boosting Complex Video Reasoning in Large Vision-Language Models through Agentic Data Synthesis](https://mlsanity.com/paper/2509.23652/) (2025)
- [DualVLA: Building a Generalizable Embodied Agent via Partial Decoupling of Reasoning and Action](https://mlsanity.com/paper/2511.22134/) (2025)
- [From Frames to Clips: Training-free Adaptive Key Clip Selection for Long-Form Video Understanding](https://mlsanity.com/paper/2510.02262/) (2025)
- [Symphony: A Cognitively-Inspired Multi-Agent System for Long-Video Understanding](https://mlsanity.com/paper/2603.17307/) (2026)
- [MemLearner: Learning to Query Context memory for Video World Models](https://mlsanity.com/paper/2606.31734/) (2026)
- [Demo-ICL: In-Context Learning for Procedural Video Knowledge Acquisition](https://mlsanity.com/paper/2602.08439/) (2026)
- [Think-as-You-See: Streaming Chain-of-Thought Reasoning for Large Vision-Language Models](https://mlsanity.com/paper/2603.02872/) (2026)
- [MME-CoF-Pro: Evaluating Reasoning Coherence in Video Generative Models with Text and Visual Hints](https://mlsanity.com/paper/2603.20194/) (2026)
- [3D-Layout-R1: Structured Reasoning for Language-Instructed Spatial Editing](https://mlsanity.com/paper/2603.22279/) (2026)
- [LAST: LeArning to Think in Space and Time for Generalist Vision-Language Models](https://mlsanity.com/paper/2511.19261/) (2025)
- [Cut to the Chase: Training-free Multimodal Summarization via Chain-of-Events](https://mlsanity.com/paper/2603.06213/) (2026)
- [Personal Visual Context Learning in Large Multimodal Models](https://mlsanity.com/paper/2605.10936/) (2026)
- [Visual Agentic Memory: Enabling Online Long Video Understanding via Online Indexing, Hierarchical Memory, and Agentic Retrieval](https://mlsanity.com/paper/2605.16481/) (2026)
- [How to Train Your Long-Context Visual Document Model](https://mlsanity.com/paper/2602.15257/) (2026)
- [Swift Sampling: Selecting Temporal Surprises via Taylor Series](https://mlsanity.com/paper/2605.22678/) (2026)
- [Audio-Visual Flamingo: Open Audio-Visual Intelligence for Long and Complex Videos](https://mlsanity.com/paper/2607.16107/) (2026)
- [TAMA: Tool-Augmented Multimodal Agent for Procedural Activity Understanding](https://mlsanity.com/paper/2510.00161/) (2025)
- [Learning Consistent Temporal Grounding between Related Tasks in Sports Coaching](https://mlsanity.com/paper/2603.18453/) (2026)
- [Why Do Vision Language Models Struggle To Recognize Human Emotions?](https://mlsanity.com/paper/2604.15280/) (2026)
- [STORM: Internalized Modeling for Spatial-Temporal Reasoning in Video-Language Models](https://mlsanity.com/paper/2605.26014/) (2026)
- [Internalized Reasoning for Long-Context Visual Document Understanding](https://mlsanity.com/paper/2604.02371/) (2026)
- [Event-Causal RAG: A Retrieval-Augmented Generation Framework for Long Video Reasoning in Complex Scenarios](https://mlsanity.com/paper/2605.06185/) (2026)
- [Zero-Shot 3D Question Answering via Hierarchical View-to-Token Transportation](https://mlsanity.com/paper/2606.03100/) (2026)
- [QSVideo: Query-Conditioned Semantic Temporal Retrieval for Video Understanding](https://mlsanity.com/paper/2607.04559/) (2026)
- [Beyond Frame Selection: Rethinking Long-Video Understanding with MLLMs](https://mlsanity.com/paper/2608.05592/) (2026)
- [Cog-VADU: A Training-Free Cognitive Reasoning Framework for Video Anomaly Detection and Understanding](https://mlsanity.com/paper/2610.01754/) (2026)

## Abstract (arXiv)

Despite recent advances in Vision-Language Models (VLMs), long-video understanding remains a challenging problem. Although state-of-the-art long-context VLMs can process around 1000 input frames, they still struggle to effectively leverage this sequence length, and succumb to irrelevant distractors within the context window. We present Temporal Chain of Thought, an inference strategy for video question-answering that curates the model's input context. We use the VLM itself to iteratively identify and extract the most relevant frames from the video, which are then used for answering. We demonstrate how leveraging more computation at inference-time to select the most relevant context leads to improvements in accuracy, in agreement with recent work on inference-time scaling of LLMs. Moreover, we achieve state-of-the-art results on 4 diverse video question-answering datasets, showing consistent improvements with 3 different VLMs. In particular, our method shines on longer videos which would not otherwise fit within the model's context window: On longer videos of more than 1 hour on LVBench, our approach using a context window of 32K outperforms the same VLM using standard inference with a 700K context window by 2.8 points.