# Diffusion Models Are Real-Time Game Engines > GameNGen learns to simulate DOOM interactively with a diffusion model instead of game code. A PPO agent first plays the game in ViZDoom, and its recorded trajectories, from random to trained play, become the training data. Stable Diffusion v1.4 is then fine-tuned to predict the next frame from the last 64 frames and actions, with noise added to context frames to keep long autoregressive rollouts stable and a fine-tuned latent decoder for detail. With 4 sampling steps it runs at 20 frames per second on one TPU-v5. The authors report next-frame PSNR of 29.43 and human raters picking the real game in 58–60% of short clips. - Authors: Dani Valevski; Yaniv Leviathan; Moab Arar; Shlomi Fruchter - Venue: ICLR 2025 - Status: Conference paper - Publication / posting date: 2025 (exact date not verified) - Date source: Publication year stated by the official ICLR proceedings; no finer publication date is asserted. The proceedings page's citation date of 2025-05-01 is not adopted, and the arXiv dates are not substituted. - First arXiv submission: August 27, 2024 - arXiv: 2408.14837 - Reviewed arXiv version: v2 - Sources reviewed: 2026-10-27 - Added to the catalog: 2026-11-03 - Original paper: - PDF: - Project: - BibTeX: - Catalog page: - Record JSON: - Chinese version: [简体中文](https://awesomerea.org/zh/papers/paper-gamengen/index.md) ## Scope and limitations A neural approximation trained on recorded gameplay of an existing game; it recovers no game logic, source code, assets or state representation and is not an exact simulation. The model sees just over 3 seconds of history, so state outside that window survives only through on-screen cues and learned heuristics, which can be wrong; for example, repeated shooting may spawn enemies. Locations and interactions the agent never explored are simulated poorly. Training used 128 TPU-v5e devices and 70M agent frames at 320×240, plus an environment-specific reward function. PSNR is a teacher-forced single-frame measure on 2,048 held-out trajectories from 5 levels; autoregressive rollouts diverge from ground truth after a few steps. Human evaluation used 10 raters on 1.6- and 3.2-second clips and 150 later 3-second clips; the authors say they could often spot the simulation within seconds. Only DOOM was tested, and the model cannot readily create new games. ## Why it fits this collection Adds learned, non-symbolic reconstruction of a game's interactive behavior from gameplay, contrasting with the collection's code-based world models and visual game replication research. ## Publication and review notes Authors and title follow the ICLR 2025 proceedings page and PDF, which list pages 73754–73776 and print “Published as a conference paper at ICLR 2025”; no DOI is listed. arXiv has two versions with the same title and authors: v1 of 2024-08-27 and v2 of 2025-04-24, whose comment reads “ICLR 2025”. The OpenReview forum P8pqeEkn1H could not be retrieved automatically because of a challenge page. The project page links the paper and videos but no code or weights, and no author-provided code release was found. The paper's reproducibility statement relies on open Stable Diffusion v1.4 weights, the open-source ViZDoom environment and documented training settings. Its reported 50 frames per second uses a distilled single-step model with some quality loss. ## Primary sources 1. 2. 3. 4. ## About this entry Reviewed against the paper and its primary sources. Cite the original authors using the linked paper or BibTeX. Original editorial content: MIT. Linked projects, papers and assets retain their own rights and licenses.