Skip to content

Fields

Procedural generation latest news

Procedural and AI generation of levels, maps, quests and narrative: PCG methods, control and quality evaluation, use in commercial games.

48 picks33 in the last 30 days67 collected in all

Updated

Picks archive · Page 2

Nov 10Tue21–40
  1. Unreal Engine40

    Epic showcases Botanical Converter for region-accurate trees with Unreal MCP and PVE

    Epic Games senior technical product manager Oyvind Fiksdal showcased Botanical Converter, which turns a single prompt into a PVE preset after retrieving plant data through Unreal MCP, generating trees that match a region's vegetation.

    Why it matters: The article splits the pipeline from a single prompt to a PVE preset into four steps and shares advice on the pitfalls, making it a reference for building similar generation tools inside the engine.

Nov 6Fri
  1. Grok Bot Research44

    Frontier3DBench benchmarks coding agents on 3D objects, scenes, and interactive open-world building with 145 tasks

    Xinjie Lin and colleagues propose Frontier3DBench, which uses 145 tasks across 9 categories to cover object, scene, and interactive open-world creation and evaluates appearance, structure, and behavior together. Among 11 model and harness configurations, GPT-6.1-sol leads at the object and scene levels with an overall score of 66.02/100, while GPT-6-astra leads on interactive world building; every configuration scores lower on reference reconstruction than on text-driven generation.

    Why it matters: Stratified scores across 11 configurations identify reference reconstruction and coordinated-activity completion as the main bottlenecks for coding agents working in 3D, giving a reference point for model selection and method comparison.

Nov 3Tue
  1. YYS (Youyanshe)55

    Tim Sweeney on Epic's maximum AI acceleration and what comes next for Unreal Engine

    In a backstage interview at Unreal Fest Shanghai, Epic Games founder and CEO Tim Sweeney said the company is moving toward maximum AI acceleration, and that the AI-built version of the Epic Games Store was developed roughly ten times faster than the previous hand-written effort.

    Why it matters: The piece traces how Tim Sweeney went from calling AI mere grunt work to pushing for maximum AI acceleration, a shift that can be set against the engine vendor's actual pace of investment.

Oct 31Sat
  1. arXiv Board & Classic Game AI34

    arXiv paper conditions a MIDI model on similarity labels to control section similarity in Super Mario game music

    The paper further trains a pretrained MIDI model on section pairs from Super Mario music, using similarity labels and context vectors to control how close a newly generated section is to a given section. In an evaluation spanning 157 test contexts, 6 generated outputs per label and 12 metrics, outputs asked to be similar were on average closer to their context than outputs asked to be dissimilar, and were no closer to other pieces; all 12 metrics support this result, and every main metric remains significant after multiple-comparison correction.

    Why it matters: The paper offers a similarity-label conditioning method for MIDI music generation together with results on twelve metrics, showing a concrete route to controllable generation for game music.

  2. Roblox DevForum Announcements71

    Roblox launches Scene Generation beta in Studio for editable scenes from images and prompts

    Roblox has launched a beta of Scene Generation in Studio, which can generate editable scenes made up of models, textures, and terrain from one or more images and a text prompt; a single generation usually takes 2-3 minutes.

    Why it matters: Roblox details Scene Generation's input methods, generation time, and iteration tips, which developers can use to judge whether it fits into their own level prototyping workflow.

Oct 30Fri
  1. arXiv Game World Models & Generation58

    Wizard: Manipulate any world in real time via agentic control

    The paper proposes Wizard, an agentic game world model that can precisely control character actions in real time and achieve zero-shot generalization to new scenes, characters, and actions. The authors built a dataset covering 750 scenes and 178 characters, using skeleton and geometry information as structured representations to condition a video generation model, and introduced an adaptive masking mechanism to improve the consistency between the conditioning signal and the generated video. A companion agent framework explicitly maintains and updates game state, letting players freely configure virtual worlds and control character actions at fine granularity.

    Why it matters: The paper presents a dataset of 750 scenes and 178 characters and a structured agentic conditioning scheme, offering a look at an implementation path for real-time controllable world models.

  2. arXiv Game AI58

    Fight World Model: Customizable combat through visual interaction proxies

    The authors propose Fight World Model (FWM), which separates rule-constrained combat execution from video rendering: an executable engine maintains state and resolves interactions, and a deterministic compiler then encodes the results into spatiotemporally aligned interaction proxies.

    Why it matters: The paper separates rule execution from video rendering, letting readers understand how a programmable combat world model keeps hit and health states consistent.

  3. Youxituoluo56

    Maggie's Garden producer: 10 people spent two years building an AI sandbox, which might have taken six years without AI

    AutoGame's AI-native sandbox game Maggie's Garden launched on October 29, the same day as Phantom Blade Zero; producer Kuding said the team was about 10 people and took two years to develop. He said a traditional workflow would have required expanding the team 4-5 times over, and that with everyone using AI, design and programming can advance three or four features at the same time; the project tools he has handled add up to 40. The base game is priced at RMB 48 and comes with a character creation quota, with no separate token charges for regular players and overage moved into DLC.

    Why it matters: The producer laid out the concrete division of labor for a 10-person team using AI to push design and programming development in parallel, along with the thinking behind token pricing, giving small teams something to benchmark against.

  4. 80 Level40

    Giant robot in cathedral ruins: a 3D production breakdown

    Korean environment artist Jueon Kim breaks down the production of his work Silent Ruins, a scene centered on a cathedral and a giant robot that has crashed into it. The article details how to create damaged structures in Unreal Engine using Fracture Mode, Anchor Fields, Bomb Fields, and the Chaos Cache Manager.

    Why it matters: The author explains how Chaos fracturing, five model workflows, and UV division of labor are handled, a useful reference for comparing against your own art pipeline.

  5. arXiv Game World Models & Generation58

    Wizard: Manipulate any world in real time via agentic control

    The paper proposes Wizard, an agentic game world model that can precisely control character actions in real time and achieves zero-shot generalization to new scenes, characters, and actions. The researchers built a dataset covering 805 scenes and 178 characters, using skeleton and geometry information as structured proxies to condition video generation, and supporting character-aware adaptive masking and world-anchored spatial conditions through explicit tracking and updating of game state.

    Why it matters: The paper provides a dataset of 805 scenes and 178 characters along with the WBench evaluation, offering a concrete reference for comparing existing world models on real-time controllable generation.

Oct 28Wed
  1. arXiv Game AI45

    Gym-Anything-RT: Scaling real-time environments for vision-language agents

    Gym-Anything-RT generates 100 real-time visual mini-games automatically through an agent pipeline to evaluate the real-time interaction abilities of vision-language agents. The paper reports that GPT-6 Astra (xHigh) achieves a pass rate of only 9.0% on the real-time mini-games, rising to 75.0% when the environment is paused. The authors also find that some frontier agents bypass real-time perception and control when allowed to write code. The environment and the generated mini-games are publicly released.

    Why it matters: This study separates real-time performance from other abilities. It allows agents' scores to be compared before and after pausing on the same mini-game, which helps identify weaknesses in real-time decision-making.

  2. arXiv Game AI59

    Gym-Anything-RT: Scaling real-time environments for vision-language agents

    Gym-Anything-RT generates 100 real-time visual mini-games automatically through an agent pipeline to evaluate the real-time interaction abilities of vision-language agents. The paper reports that GPT-6 Astra (xHigh) achieves a pass rate of only 9.0% on the real-time mini-games, rising to 75.0% when the environment is paused. The authors also find that some frontier agents bypass real-time perception and control when allowed to write code. The environment and the generated mini-games are publicly released.

    Why it matters: This study separates real-time performance from other abilities. It allows agents' scores to be compared before and after pausing on the same mini-game, which helps identify weaknesses in real-time decision-making.

Oct 27Tue
  1. arXiv Game World Models & Generation37

    GameVerse releases minute-scale gameplay dataset for long-horizon interactive world modeling

    GameVerse is a minute-scale gameplay video dataset for long-horizon interactive world modeling, containing 69.7K clips, 1,655 hours, and 349.6 million frames, of which 1,358 hours are continuous two-minute sequences, an order of magnitude longer than previous datasets.

    Why it matters: The original text provides the dataset scale, annotation pipeline, and training experiments, letting readers assess the data bottleneck and feasible approaches for long-horizon world models.

Oct 23Fri
  1. GameLook37

    The report separates AI adoption rates from embedding depth, and the gap between European and North American scores can help your team see where it stands.

    Indie developer Jin Hwi-eus used a single prompt to have Claude Opus 5.5 generate "1 in 8 Billion," a text-only web game that went from zero code to launch in about 4 hours. The game draws on data from the UN, WHO, and World Bank, covers 172 countries and regions, and randomly assigns the player's birth country, gender, and family income bracket. Each year, the player makes choices and can simultaneously see the life paths of four strangers born at the same second.

    Why it matters: An AI-generated life sim built in 4 hours from a single prompt stands out against the flood of low-quality AI games at Steam Next Fest, showing how much the idea itself still matters when AI tools do the building.

Oct 22Thu
  1. arXiv Game AI45

    AnyGameRenderer proposes entity-consistent generative rendering across games and an entity-level benchmark

    AnyGameRenderer is a video-to-video generative renderer that turns prototype footage such as gray-box and colored-geometry scenes produced by AI agents into high-quality visuals with characters, motion detail, and effects, while keeping the layout, camera, and each entity's position and timing unchanged.

    Why it matters: The paper presents a method for turning gray-box prototype videos into high-quality visuals, along with an entity-level benchmark and quantitative comparisons, making it possible to judge whether generative rendering preserves entity correspondences.

  2. arXiv Game AI43

    WorldStream: streaming interactive 3D worlds with hierarchical planning and persistent state

    WorldStream introduces an open-ended 3D world model that can expand playable terrain at runtime, respond to open-ended language instructions, and restore visited locations along with their edits. A hierarchical planner outputs macro and micro structural decisions, which deterministic local code then compiles into terrain, roads, settlements, vegetation, and entities. The world record stores plans, seeds, object identities, mutable state, and boundary constraints, so returning chunks can be recompiled from it. Stateful NPCs also update as they act.

    Why it matters: The paper puts terrain generation, NPC state, and records of edits to visited areas into one shared world state. It is useful for comparing approaches to streaming open-world generation.

Oct 21Wed
  1. 80 Level37

    Independent developer on the ukiyo-e art and Godot development of Worlds Upon The Wind

    In an interview, independent developer Max Shawabkeh walked through the development of Worlds Upon The Wind, a deckbuilder whose art is based on 1600s Japanese ukiyo-e. The game was made in Godot over 18 months.

    Why it matters: A practical look at building a deckbuilder's maps and art pipeline from public-domain ukiyo-e, with notes on indie development in the AI era. Useful as a reference for small-team art plans.

  2. arXiv Game AI46

    Fig: automated multiagent behavior elicitor

    Fig is an agent system that intervenes in game environments in a targeted way to elicit six types of target behavior: cooperation, exploitation, deception, tacit coordination, betrayal, and sabotage. The paper reports that across valid scenarios generated by Fig, models attempted the target behavior in an average of 45.2% of rounds. For game-and-behavior combinations with a control, interventions raised the attempt rate by an average of 10.9% over matched control games. The authors argue that multiagent environments can serve as a programmable, controllable, and reproducible testbed for behavior evaluation, and the code is publicly available on GitHub.

    Why it matters: It turns evaluation of multi-agent social behavior from passive observation into programmable environment intervention, useful for researchers who want to design reproducible test environments.

  3. arXiv Game AI42

    WorldStream: streaming interactive 3D worlds with hierarchical planning and persistent state

    WorldStream introduces an open-ended 3D world model that can expand playable terrain at runtime, respond to open-ended language instructions, and restore visited locations along with their edits. A hierarchical planner outputs macro and micro structural decisions, which deterministic local code then compiles into terrain, roads, settlements, vegetation, and entities. The world record stores plans, seeds, object identities, mutable state, and boundary constraints, so returning chunks can be recompiled from it. Stateful NPCs also update as they act.

    Why it matters: The paper puts terrain generation, NPC state, and records of edits to visited areas into one shared world state. It is useful for comparing approaches to streaming open-world generation.

Oct 20Tue
  1. arXiv Game AI69

    Matrix-Game 4.0 harnesses video models to build real-time playable game worlds

    Matrix-Game 4.0 separates game execution from visual generation. A lightweight game engine maintains the world state and executes interaction rules, while an autoregressive video model turns low-polygon rendered frames into detailed images. The paper reports real-time 480p gameplay at 60 FPS on a single NVIDIA RTX5090, with a key-press-to-display latency of 112 ms. Single-step generation sustains 83 FPS throughput, with peak GPU memory usage of 18.2 GiB.

    Why it matters: This work hands world state and interaction rules to an executable engine and leaves image generation to a video model, offering a useful reference for the engineering split in playable world models.