# QuantWM: Temporally Consistent 2-Bit KV Cache Quantization for World Models and Video Generation

arXiv: https://arxiv.org/abs/2609.26425 · HTML: https://mlsanity.com/paper/2609.26425/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Video-generation world models and simulators, one of 11 directions in World models.
- Builds on 34 of its 39 references that are in the map.
- First posted on arXiv 22 Sep 2026.

## Placed in

- World models / [Video-generation world models and simulators](https://mlsanity.com/wm/video-generation-world-models-and-simulators/)
- Efficient LLM inference and serving / [KV cache quantization](https://mlsanity.com/inference/kv-cache-quantization/)

## Builds on (most cited first, 30 of 34)

- [CogVideoX: Text-to-Video Diffusion Models with An Expert Transformer](https://mlsanity.com/paper/2408.06072/) (2024)
- [GPTQ: Accurate Post-Training Quantization for Generative Pre-trained Transformers](https://mlsanity.com/paper/2210.17323/) (2022)
- [VBench: Comprehensive Benchmark Suite for Video Generative Models](https://mlsanity.com/paper/2311.17982/) (2023)
- [Make-A-Video: Text-to-Video Generation without Text-Video Data](https://mlsanity.com/paper/2209.14792/) (2022)
- [SmoothQuant: Accurate and Efficient Post-Training Quantization for Large Language Models](https://mlsanity.com/paper/2211.10438/) (2022)
- [AWQ: Activation-aware Weight Quantization for LLM Compression and Acceleration](https://mlsanity.com/paper/2306.00978/) (2023)
- [Open-Sora: Democratizing Efficient Video Production for All](https://mlsanity.com/paper/2412.20404/) (2024)
- [Movie Gen: A Cast of Media Foundation Models](https://mlsanity.com/paper/2410.13720/) (2024)
- [KIVI: A Tuning-Free Asymmetric 2bit Quantization for KV Cache](https://mlsanity.com/paper/2402.02750/) (2024)
- [Self Forcing: Bridging the Train-Test Gap in Autoregressive Video Diffusion](https://mlsanity.com/paper/2506.08009/) (2025)
- [Up or Down? Adaptive Rounding for Post-Training Quantization](https://mlsanity.com/paper/2004.10568/) (2020)
- [KVQuant: Towards 10 Million Context Length LLM Inference with KV Cache Quantization](https://mlsanity.com/paper/2401.18079/) (2024)
- [OmniQuant: Omnidirectionally Calibrated Quantization for Large Language Models](https://mlsanity.com/paper/2308.13137/) (2023)
- [LLM-QAT: Data-Free Quantization Aware Training for Large Language Models](https://mlsanity.com/paper/2305.17888/) (2023)
- [Diffusion for World Modeling: Visual Details Matter in Atari](https://mlsanity.com/paper/2405.12399/) (2024)
- [Diffusion Models Are Real-Time Game Engines](https://mlsanity.com/paper/2408.14837/) (2024)
- [LongLive: Real-time Interactive Long Video Generation](https://mlsanity.com/paper/2509.22622/) (2025)
- [Rolling Forcing: Autoregressive Long Video Diffusion in Real Time](https://mlsanity.com/paper/2509.25161/) (2025)
- [QServe: W4A8KV4 Quantization and System Co-design for Efficient LLM Serving](https://mlsanity.com/paper/2405.04532/) (2024)
- [Scaling Autoregressive Video Models](https://mlsanity.com/paper/1906.02634/) (2019)
- [Matrix-game 2.0: An open-source real-time and streaming interactive world model](https://mlsanity.com/paper/2508.13009/) (2025)
- [Causal Forcing: Autoregressive Diffusion Distillation Done Right for High-Quality Real-Time Interactive Video Generation](https://mlsanity.com/paper/2602.02214/) (2026)
- [WorldPlay: Towards Long-Term Geometric Consistency for Real-Time Interactive World Modeling](https://mlsanity.com/paper/2512.14614/) (2025)
- [WorldMem: Long-term Consistent World Simulation with Memory](https://mlsanity.com/paper/2504.12369/) (2025)
- [Matrix-Game 3.0: Real-Time and Streaming Interactive World Model with Long-Horizon Memory](https://mlsanity.com/paper/2604.08995/) (2026)
- [KV Cache is 1 Bit Per Channel: Efficient Large Language Model Inference with Coupled Quantization](https://mlsanity.com/paper/2405.03917/) (2024)
- [SliM-LLM: Salience-Driven Mixed-Precision Quantization for Large Language Models](https://mlsanity.com/paper/2405.14917/) (2024)
- [Infinite Worlds with Versatile Interactions](https://mlsanity.com/paper/2607.07534/) (2026)
- [CoLA: Compute-Efficient Pre-Training of LLMs via Low-Rank Activation](https://mlsanity.com/paper/2502.10940/) (2025)
- [CommVQ: Commutative Vector Quantization for KV Cache Compression](https://mlsanity.com/paper/2506.18879/) (2025)

## Built on by (most cited first, 1 of 1)

- [CORAL-KV: Compact Cross-Layer Prediction and Residual Coding for Autoregressive Video KV Caches](https://mlsanity.com/paper/2610.12514/) (2026)

## Abstract (arXiv)

Video world models achieve long-range temporal consistency by storing KV cache during generation, but the growing cache makes KV cache memory a major deployment bottleneck, which motivates low-bit quantization study for efficiency. Existing 2-bit KV cache quantization methods can achieve nearly lossless performance on VBench, however, when applied to video world models, we find they still cause severe temporal flickering and visual degradation. Meanwhile, deeper investigates show that Key quantization produces smaller reconstruction errors than Value, but surprisingly leads to larger output degradation. We trace this discrepancy to attention in video world models: Key perturbations can change the attention logits, and shift the temporal-spatial tokens selected by Queries. These observations motivate us to preserve attention logits and temporal-spatial token selection during KV cache quantization. To address this issue, we present QuantWM, a training-free 2-bit KV cache quantization framework for video world models. QuantWM introduces two complementary techniques to mitigate the attention shifts. Firstly, quantization-sensitivity-aware clustering (QSAC) jointly considers historical Query sensitivity and residual ranges to select INT2-friendly Key centroids, which reduces quantization errors in channels that are more critical to attention. In addition, principal-subspace attention compensation (PSAC) restores the remaining Key errors along the dominant Query subspace using low-rank projections, which provides a direct and efficient correction to stabilize attention logits. Experiments on LingBot-World-v2, HY-World 1.5, Matrix-Game-2, Longcat-Video and Causal-Forcing demonstrate that QuantWM significantly improves visual quality and temporal consistency, while outperforming existing methods across benchmarks with up to 6.20 KV cache memory compression and limited additional overhead.