# LatSearch: Latent Reward-Guided Search for Faster Inference-Time Scaling in Video Diffusion

arXiv: https://arxiv.org/abs/2603.14526 · HTML: https://mlsanity.com/paper/2603.14526/

## Key facts

- Cited by 10 papers in the map (data through 14 Oct 2026).
- Placed in Diffusion reasoning and test-time scaling, one of 17 directions in LLM reasoning.
- Builds on 28 of its 64 references that are in the map.
- First posted on arXiv 15 Mar 2026.

## Placed in

- LLM reasoning / [Diffusion reasoning and test-time scaling](https://mlsanity.com/lr/diffusion-reasoning-and-test-time-scaling/)
- Diffusion models / [Video generation](https://mlsanity.com/diffusion/video-generation/)
- LLM alignment / RLHF / preference optimisation / [Diffusion and multimodal model alignment](https://mlsanity.com/alignment/diffusion-and-multimodal-model-alignment/)

## Builds on (most cited first, 28 of 28)

- [Denoising Diffusion Implicit Models](https://mlsanity.com/paper/2010.02502/) (2020)
- [Classifier-Free Diffusion Guidance](https://mlsanity.com/paper/2207.12598/) (2022)
- [Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution](https://mlsanity.com/paper/2409.12191/) (2024)
- [Wan: Open and Advanced Large-Scale Video Generative Models](https://mlsanity.com/paper/2503.20314/) (2025)
- [Stable Video Diffusion: Scaling Latent Video Diffusion Models to Large Datasets](https://mlsanity.com/paper/2311.15127/) (2023)
- [HunyuanVideo: A Systematic Framework For Large Video Generative Models](https://mlsanity.com/paper/2412.03603/) (2024)
- [Cosmos World Foundation Model Platform for Physical AI](https://mlsanity.com/paper/2501.03575/) (2025)
- [ModelScope Text-to-Video Technical Report](https://mlsanity.com/paper/2308.06571/) (2023)
- [VideoCrafter1: Open Diffusion Models for High-Quality Video Generation](https://mlsanity.com/paper/2310.19512/) (2023)
- [VBench-2.0: Advancing Video Generation Benchmark Suite for Intrinsic Faithfulness](https://mlsanity.com/paper/2503.21755/) (2025)
- [A General Framework for Inference-time Scaling and Steering of Diffusion Models](https://mlsanity.com/paper/2501.06848/) (2025)
- [Autoregressive Video Generation without Vector Quantization](https://mlsanity.com/paper/2412.14169/) (2024)
- [Long-Context Autoregressive Video Modeling with Next-Frame Prediction](https://mlsanity.com/paper/2503.19325/) (2025)
- [Golden Noise for Diffusion Models: A Learning Framework](https://mlsanity.com/paper/2411.09502/) (2024)
- [Not All Noises Are Created Equally:Diffusion Noise Selection and Optimization](https://mlsanity.com/paper/2407.14041/) (2024)
- [Control-A-Video: Controllable Text-to-Video Diffusion Models with Motion Prior and Reward Feedback Learning](https://mlsanity.com/paper/2305.13840/) (2023)
- [Video-T1: Test-Time Scaling for Video Generation](https://mlsanity.com/paper/2503.18942/) (2025)
- [Scaling Image and Video Generation via Test-Time Evolutionary Search](https://mlsanity.com/paper/2505.17618/) (2025)
- [DOLLAR: Few-Step Video Generation via Distillation and Latent Reward Optimization](https://mlsanity.com/paper/2412.15689/) (2024)
- [Training-Free Motion-Guided Video Generation with Enhanced Temporal Consistency Using Motion Consistency Loss](https://mlsanity.com/paper/2501.07563/) (2025)
- [ScalingNoise: Scaling Inference-Time Search for Generating Infinite Videos](https://mlsanity.com/paper/2503.16400/) (2025)
- [Enhance-A-Video: Better Generated Video for Free](https://mlsanity.com/paper/2502.07508/) (2025)
- [Model Already Knows the Best Noise: Bayesian Active Noise Selection via Attention in Video Diffusion Model](https://mlsanity.com/paper/2505.17561/) (2025)
- [SwiftVideo: A Unified Framework for Few-Step Video Generation through Trajectory-Distribution Alignment](https://mlsanity.com/paper/2508.06082/) (2025)
- [Aligning Anime Video Generation with Human Feedback](https://mlsanity.com/paper/2504.10044/) (2025)
- [Tuning-Free Long Video Generation via Global-Local Collaborative Diffusion](https://mlsanity.com/paper/2501.05484/) (2025)
- [CoCA: Step-level Reward for Free in RL-based T2I Diffusion Model Fine-tuning](https://mlsanity.com/paper/2505.19196/) (2025)
- [TokensGen: Harnessing Condensed Tokens for Long Video Generation](https://mlsanity.com/paper/2507.15728/) (2025)

## Built on by (most cited first, 10 of 10)

- [Relax Forcing: Relaxed KV-Memory for Consistent Long Video Generation](https://mlsanity.com/paper/2603.21366/) (2026)
- [Stream-T1: Test-Time Scaling for Streaming Video Generation](https://mlsanity.com/paper/2605.04461/) (2026)
- [Inference-Time Scaling for Joint Audio-Video Generation](https://mlsanity.com/paper/2606.03183/) (2026)
- [CachedSearch: Training-Free Cached Exploration for Test-Time Search in Video Diffusion](https://mlsanity.com/paper/2607.23159/) (2026)
- [Stitched Value Model for Diffusion Alignment](https://mlsanity.com/paper/2605.19804/) (2026)
- [Flat-Consensus Diffusion for Robust Data Reshaping under Noisy Evaluator](https://mlsanity.com/paper/2609.32696/) (2026)
- [Test-Time Scaling for Video Diffusion Models via Diagnosis-Guided Candidate Recycling](https://mlsanity.com/paper/2608.29322/) (2026)
- [AutoRef: Harness Optimization for Agentic Multi-Reference Image Generation](https://mlsanity.com/paper/2609.35530/) (2026)
- [VIF-Bench: Evaluating Visual Instruction Following in Multi-Reference Image Generation](https://mlsanity.com/paper/2609.37709/) (2026)
- [PreCoG: Preview-Edit for Efficient Video Correction in Generation-Time](https://mlsanity.com/paper/2610.13174/) (2026)

## Abstract (arXiv)

The recent success of inference-time scaling in large language models has inspired similar explorations in video diffusion. In particular, motivated by the existence of "golden noise" that enhances video quality, prior work has attempted to improve inference by optimising or searching for better initial noise. However, these approaches have notable limitations: they either rely on priors imposed at the beginning of noise sampling or on rewards evaluated only on the denoised and decoded videos. This leads to error accumulation, delayed and sparse reward signals, and prohibitive computational cost, which prevents the use of stronger search algorithms. Crucially, stronger search algorithms are precisely what could unlock substantial gains in controllability, sample efficiency and generation quality for video diffusion, provided their computational cost can be reduced. To fill in this gap, we enable efficient inference-time scaling for video diffusion through latent reward guidance, which provides intermediate, informative and efficient feedback along the denoising trajectory. We introduce a latent reward model that scores partially denoised latents at arbitrary timesteps with respect to visual quality, motion quality, and text alignment. Building on this model, we propose LatSearch, a novel inference-time search mechanism that performs Reward-Guided Resampling and Pruning (RGRP). In the resampling stage, candidates are sampled according to reward-normalised probabilities to reduce over-reliance on the reward model. In the pruning stage, applied at the final scheduled step, only the candidate with the highest cumulative reward is retained, improving both quality and efficiency. We evaluate LatSearch on the VBench-2.0 benchmark and demonstrate that it consistently improves video generation across multiple evaluation dimensions compared to the baseline Wan2.1 model.