# InfLVG: Reinforce Inference-Time Consistent Long Video Generation with GRPO

arXiv: https://arxiv.org/abs/2505.17574 · HTML: https://mlsanity.com/paper/2505.17574/

## Key facts

- Cited by 15 papers in the map (data through 14 Oct 2026).
- Placed in Multimodal reasoning, one of 17 directions in LLM reasoning.
- Builds on 20 of its 51 references that are in the map.
- First posted on arXiv 23 May 2025.

## Placed in

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

## Builds on (most cited first, 20 of 20)

- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [Denoising Diffusion Implicit Models](https://mlsanity.com/paper/2010.02502/) (2020)
- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [Flow Matching for Generative Modeling](https://mlsanity.com/paper/2210.02747/) (2022)
- [Qwen2.5-VL Technical Report](https://mlsanity.com/paper/2502.13923/) (2025)
- [CogVideoX: Text-to-Video Diffusion Models with An Expert Transformer](https://mlsanity.com/paper/2408.06072/) (2024)
- [Training Diffusion Models with Reinforcement Learning](https://mlsanity.com/paper/2305.13301/) (2023)
- [Human Preference Score v2: A Solid Benchmark for Evaluating Human Preferences of Text-to-Image Synthesis](https://mlsanity.com/paper/2306.09341/) (2023)
- [Gemini: A Family of Highly Capable Multimodal Models](https://mlsanity.com/paper/2312.11805/) (2023)
- [Vision-R1: Incentivizing Reasoning Capability in Multimodal Large Language Models](https://mlsanity.com/paper/2503.06749/) (2025)
- [InternVid: A Large-scale Video-Text Dataset for Multimodal Understanding and Generation](https://mlsanity.com/paper/2307.06942/) (2023)
- [From Slow Bidirectional to Fast Autoregressive Video Diffusion Models](https://mlsanity.com/paper/2412.07772/) (2024)
- [Improving Video Generation with Human Feedback](https://mlsanity.com/paper/2501.13918/) (2025)
- [VideoScore: Building Automatic Metrics to Simulate Fine-grained Human Feedback for Video Generation](https://mlsanity.com/paper/2406.15252/) (2024)
- [FreeNoise: Tuning-Free Longer Video Diffusion via Noise Rescheduling](https://mlsanity.com/paper/2310.15169/) (2023)
- [FIFO-Diffusion: Generating Infinite Videos from Text without Training](https://mlsanity.com/paper/2405.11473/) (2024)
- [Scaling Diffusion Transformers to 16 Billion Parameters](https://mlsanity.com/paper/2407.11633/) (2024)
- [Ouroboros-Diffusion: Exploring Consistent Content Generation in Tuning-free Long Video Diffusion](https://mlsanity.com/paper/2501.09019/) (2025)
- [GRADEO: Towards Human-Like Evaluation for Text-to-Video Generation via Multi-Step Reasoning](https://mlsanity.com/paper/2503.02341/) (2025)
- [Choosing to Rank](https://mlsanity.com/paper/1809.05139/) (2018)

## Built on by (most cited first, 15 of 15)

- [The Landscape of Agentic Reinforcement Learning for LLMs: A Survey](https://mlsanity.com/paper/2509.02547/) (2025)
- [Self-Forcing++: Towards Minute-Scale High-Quality Video Generation](https://mlsanity.com/paper/2510.02283/) (2025)
- [A Survey of Reinforcement Learning for Large Reasoning Models](https://mlsanity.com/paper/2509.08827/) (2025)
- [Reward Forcing: Efficient Streaming Video Generation with Rewarded Distribution Matching Distillation](https://mlsanity.com/paper/2512.04678/) (2025)
- [LoL: Longer than Longer, Scaling Video Generation to Hour](https://mlsanity.com/paper/2601.16914/) (2026)
- [LoViC: Efficient Long Video Generation with Context Compression](https://mlsanity.com/paper/2507.12952/) (2025)
- [Integrating Reinforcement Learning with Visual Generative Models: Foundations and Advances](https://mlsanity.com/paper/2508.10316/) (2025)
- [Reinforcement Learning for Large Model: A Survey](https://mlsanity.com/paper/2508.08189/) (2025)
- [Advancing Narrative Long Video Generation via Training-Free Identity-Aware Memory](https://mlsanity.com/paper/2605.18733/) (2026)
- [FadeMem: Distance-Aware Memory Consolidation for Autoregressive Video Diffusion](https://mlsanity.com/paper/2606.10671/) (2026)
- [Rethinking Reward Signals in Video GRPO: When Scores Become Targets](https://mlsanity.com/paper/2511.19356/) (2025)
- [SlotMemory: Object-Centric KV Memory for Streaming Long-Video Generation](https://mlsanity.com/paper/2605.31033/) (2026)
- [SpecLoR: Spectral Lookahead Rectification for Motion-Coherent Text-to-Video Generation](https://mlsanity.com/paper/2606.11969/) (2026)
- [Compositional Video Generation via Inference-Time Guidance](https://mlsanity.com/paper/2605.14988/) (2026)
- [From Evaluation to Enhancement: Benchmarking and Improving Think-with-Video Reasoning for Video Generative Models](https://mlsanity.com/paper/2609.11242/) (2026)

## Abstract (arXiv)

Recent advances in text-to-video generation, particularly with autoregressive models, have enabled the synthesis of high-quality videos depicting individual scenes. However, extending these models to generate long, cross-scene videos remains a significant challenge. As the context length grows during autoregressive decoding, computational costs rise sharply, and the model's ability to maintain consistency and adhere to evolving textual prompts deteriorates. We introduce InfLVG, an inference-time framework that enables coherent long video generation without requiring additional long-form video data. InfLVG leverages a learnable context selection policy, optimized via Group Relative Policy Optimization (GRPO), to dynamically identify and retain the most semantically relevant context throughout the generation process. Instead of accumulating the entire generation history, the policy ranks and selects the top-K most contextually relevant tokens, allowing the model to maintain a fixed computational budget while preserving content consistency and prompt alignment. To optimize the policy, we design a hybrid reward function that jointly captures semantic alignment, cross-scene consistency, and artifact reduction. To benchmark performance, we introduce the Cross-scene Video Benchmark (CsVBench) along with an Event Prompt Set (EPS) that simulates complex multi-scene transitions involving shared subjects and varied actions/backgrounds. Experimental results show that InfLVG can extend video length by up to 9×, achieving strong consistency and semantic fidelity across scenes. Our code is available at https://github.com/MAPLE-AIGC/InfLVG.