# ReasonAct: Progressive Training for Fine-Grained Video Reasoning in Small Models

arXiv: https://arxiv.org/abs/2508.01533 · HTML: https://mlsanity.com/paper/2508.01533/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Placed in Multimodal reasoning, one of 17 directions in LLM reasoning.
- Builds on 23 of its 48 references that are in the map.
- First posted on arXiv 3 Aug 2025.

## Placed in

- LLM reasoning / [Multimodal reasoning](https://mlsanity.com/lr/multimodal-reasoning/)

## Builds on (most cited first, 23 of 23)

- [Training Verifiers to Solve Math Word Problems](https://mlsanity.com/paper/2110.14168/) (2021)
- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [Measuring Mathematical Problem Solving With the MATH Dataset](https://mlsanity.com/paper/2103.03874/) (2021)
- [PaLM: Scaling Language Modeling with Pathways](https://mlsanity.com/paper/2204.02311/) (2022)
- [Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge](https://mlsanity.com/paper/1803.05457/) (2018)
- [Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback](https://mlsanity.com/paper/2204.05862/) (2022)
- [InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning](https://mlsanity.com/paper/2305.06500/) (2023)
- [VQA: Visual Question Answering](https://mlsanity.com/paper/1505.00468/) (2015)
- [Can a Suit of Armor Conduct Electricity? A New Dataset for Open Book Question Answering](https://mlsanity.com/paper/1809.02789/) (2018)
- [SlowFast Networks for Video Recognition](https://mlsanity.com/paper/1812.03982/) (2018)
- [CommonsenseQA: A Question Answering Challenge Targeting Commonsense Knowledge](https://mlsanity.com/paper/1811.00937/) (2018)
- [Is Space-Time Attention All You Need for Video Understanding?](https://mlsanity.com/paper/2102.05095/) (2021)
- [TSM: Temporal Shift Module for Efficient Video Understanding](https://mlsanity.com/paper/1811.08383/) (2018)
- [Multimodal Chain-of-Thought Reasoning in Language Models](https://mlsanity.com/paper/2302.00923/) (2023)
- [Large Language Models Can Self-Improve](https://mlsanity.com/paper/2210.11610/) (2022)
- [Video-R1: Reinforcing Video Reasoning in MLLMs](https://mlsanity.com/paper/2503.21776/) (2025)
- [Temporal Relational Reasoning in Videos](https://mlsanity.com/paper/1711.08496/) (2017)
- [COIN: A Large-scale Dataset for Comprehensive Instructional Video Analysis](https://mlsanity.com/paper/1903.02874/) (2019)
- [Temporal Action Detection with Structured Segment Networks](https://mlsanity.com/paper/1704.06228/) (2017)
- [ReClor: A Reading Comprehension Dataset Requiring Logical Reasoning](https://mlsanity.com/paper/2002.04326/) (2020)
- [Video Super Resolution Based on Deep Learning: A Comprehensive Survey](https://mlsanity.com/paper/2007.12928/) (2020)
- [OmniVec2 -- A Novel Transformer based Network for Large Scale Multimodal and Multitask Learning](https://mlsanity.com/paper/2507.13364/) (2025)
- [TemporalVLM: Video LLMs for Temporal Reasoning in Long Videos](https://mlsanity.com/paper/2412.02930/) (2024)

## Built on by (most cited first, 4 of 4)

- [RAM: Recover Any 3D Human Motion in-the-Wild](https://mlsanity.com/paper/2603.19929/) (2026)
- [Video-LMM Post-Training: A Deep Dive into Video Reasoning with Large Multimodal Models](https://mlsanity.com/paper/2510.05034/) (2025)
- [MultiPress: A Multi-Agent Framework for Interpretable Multimodal News Classification](https://mlsanity.com/paper/2604.03586/) (2026)
- [SAVOR: Self-Aware Visual Grounding via Confidence-Calibrated Reinforcement Learning for Multimodal Hallucination Mitigation](https://mlsanity.com/paper/2609.16601/) (2026)

## Abstract (arXiv)

While recent multimodal models have shown progress in vision-language tasks, small-scale variants still struggle with the fine-grained temporal reasoning required for video understanding. We introduce ReasonAct, a method that enhances video reasoning in smaller models through a three-stage training process: first building a foundation with text-only reasoning, then fine-tuning on video, and finally refining with temporal-aware reinforcement learning. We build upon Temporal Group Relative Policy Optimization (T-GRPO) by incorporating temporal consistency modeling into policy optimization. We also propose a biomechanically-motivated sub-action decomposition mechanism that provides graduated rewards for constituent action phases. Through experiments on HMDB51, UCF-101, and Kinetics-400, our 3B-parameter model achieves 67.2%, 94.1%, and 78.9% accuracy respectively, demonstrating improvements of 17.9, 15.8, and 12.3 points over baselines. Ablation studies validate that our progressive training enables smaller models to achieve competitive video reasoning performance while maintaining computational efficiency.