# Scaling Video Generation for Reasoning: At What Cost?

arXiv: https://arxiv.org/abs/2609.36599 · HTML: https://mlsanity.com/paper/2609.36599/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 30 of its 37 references that are in the map.
- First posted on arXiv 29 Sep 2026.

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

- [Scaling Laws for Neural Language Models](https://mlsanity.com/paper/2001.08361/) (2020)
- [Scalable Diffusion Models with Transformers](https://mlsanity.com/paper/2212.09748/) (2022)
- [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://mlsanity.com/paper/2104.09864/) (2021)
- [Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flow](https://mlsanity.com/paper/2209.03003/) (2022)
- [Training Compute-Optimal Large Language Models](https://mlsanity.com/paper/2203.15556/) (2022)
- [Wan: Open and Advanced Large-Scale Video Generative Models](https://mlsanity.com/paper/2503.20314/) (2025)
- [RAFT: Recurrent All-Pairs Field Transforms for Optical Flow](https://mlsanity.com/paper/2003.12039/) (2020)
- [VBench: Comprehensive Benchmark Suite for Video Generative Models](https://mlsanity.com/paper/2311.17982/) (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)
- [Representation Alignment for Generation: Training Diffusion Transformers Is Easier Than You Think](https://mlsanity.com/paper/2410.06940/) (2024)
- [Deep Learning Scaling is Predictable, Empirically](https://mlsanity.com/paper/1712.00409/) (2017)
- [LTX-Video: Realtime Video Latent Diffusion](https://mlsanity.com/paper/2501.00103/) (2024)
- [Smarter, Better, Faster, Longer: A Modern Bidirectional Encoder for Fast, Memory Efficient, and Long Context Finetuning and Inference](https://mlsanity.com/paper/2412.13663/) (2024)
- [World Action Models are Zero-shot Policies](https://mlsanity.com/paper/2602.15922/) (2026)
- [Learning Interactive Real-World Simulators](https://mlsanity.com/paper/2310.06114/) (2023)
- [Cosmos Policy: Fine-Tuning Video Models for Visuomotor Control and Planning](https://mlsanity.com/paper/2601.16163/) (2026)
- [VideoPhy: Evaluating Physical Commonsense for Video Generation](https://mlsanity.com/paper/2406.03520/) (2024)
- [World Simulation with Video Foundation Models for Physical AI](https://mlsanity.com/paper/2511.00062/) (2025)
- [Beyond Chinchilla-Optimal: Accounting for Inference in Language Model Scaling Laws](https://mlsanity.com/paper/2401.00448/) (2023)
- [WorldModelBench: Judging Video Generation Models As World Models](https://mlsanity.com/paper/2502.20694/) (2025)
- [Video as the New Language for Real-World Decision Making](https://mlsanity.com/paper/2402.17139/) (2024)
- [Scaling Laws For Diffusion Transformers](https://mlsanity.com/paper/2410.08184/) (2024)
- [A Very Big Video Reasoning Suite](https://mlsanity.com/paper/2602.20159/) (2026)
- [Towards Precise Scaling Laws for Video Diffusion Transformers](https://mlsanity.com/paper/2411.17470/) (2024)
- [MBench: A Comprehensive Benchmark on Memory Capability for Video World Models](https://mlsanity.com/paper/2606.00793/) (2026)
- [MemoBench: Benchmarking World Modeling in Dynamically Changing Environments](https://mlsanity.com/paper/2606.27537/) (2026)
- [VBVR-Pro: A Scalable and Verifiable Suite for Native Visual Reasoning](https://mlsanity.com/paper/2608.26105/) (2026)
- [Can Video World Models Track Unobserved World States?](https://mlsanity.com/paper/2608.30692/) (2026)
- [Abra: Scaling Diffusion Image Training](https://mlsanity.com/paper/2608.17286/) (2026)

## Abstract (arXiv)

We study whether scaling video generation enables models to reason about hidden information from the past frames, and at what computational cost. Our controlled benchmark requires predicting nine prescribed moves of an initially solved 2x2x2 Rubik's Cube from a fixed view of three faces. Correct predictions require inferring how actions change hidden states, and the simulator provides exact ground truth for evaluation. Models learn plausible cube geometry early, while correct sticker configurations require substantially more training. Although validation MSE follows approximate power-law scaling, lower MSE loss does not reliably indicate downstream reasoning capabilities. Smaller autoregressive models achieve higher state accuracy with limited compute, while larger models reach higher accuracy after more training. At roughly 0.1 PF-days, the 70M-parameter model correctly predicts the visible sticker configuration in 44.6% of post-action frames, compared with 0.3% for the 1B model, which reaches 83.7% at 3.14 PF-days. Symbolic state supervision raises the 20M model's frame accuracy from 31.1% to 67.3% at the same training-data budget, suggesting that learning representations of state changes can complement scaling.