#Game agents

Mine Odyssey proposes an agentic spatial intelligence benchmark with 180 tasks, rebuilding 30 real locations in Minecraft

Yuxuan Cao and others propose Mine Odyssey, an agentic spatial intelligence benchmark that rebuilds 30 real locations across 20 countries and regions on five continents in Minecraft, with 180 natural language tasks, 20 of them outdoor scenes and 10 indoor scenes.

Can AI agents learn their way to the top? AAArena evaluates heuristic learning in a long-running game competition

The paper proposes the AAArena benchmark, using 12 real adversarial games and 1,920 archived human programs to evaluate agents' learning ability in long-running competitions. Among the evaluated model and tool configurations, Opus5.5 with Claude Code took 6 gold medals, and no configuration could take the remaining 6 human ladders.

AgentGarten: a code world framework for continually evolving agents

The paper proposes the AgentGarten framework, which couples simulators and game engines to a shared neural renderer to build real-time interactive virtual environments. The simulation backend maintains persistent world state and executes program-defined interaction rules, while the renderer generates visual observations from structured conditions exported through a unified interface; the neural renderer is adapted from a pretrained video model to take geometry-conditioned input, is distilled with the proposed Adversarial Forcing, and has its inference optimized for real-time interaction.

Epic launches Unreal MCP for Unreal Engine with official support for agentic LLM tools in the editor

Epic launched the Unreal Model Context Protocol (MCP) for Unreal Engine, officially supporting the connection of commonly used agentic LLM tools directly into Unreal Engine. The protocol is built on the open MCP standard, letting agents perform operations in projects through the capabilities exposed by each editor. Developers can use compatible agentic coding tools, or build integrations, skills, and workflows around their own way of creating; documentation is at https://epic.gm/ue-mcp-docs.

A First Look at Unreal MCP in Unreal Engine 5.8

Epic has introduced Unreal MCP (Unreal Model Context Protocol) in Unreal Engine 5.8, giving agentic LLM tools an officially supported way to connect.

Same story as the pick “Epic launches Unreal MCP for Unreal Engine with official support for agentic LLM tools in the editor”

Kuration SDK: Addressing the Virtual2Real Gap in World Models via Data Curation

The paper proposes and open-sources along with it Kuration SDK, a general-purpose physical AI data curation toolkit that uses data curation to address the Virtual2Real gap in training action-conditioned world models. The authors train and evaluate diffusion world models on CounterStrike gameplay data, confirming that metrics such as FVD, LPIPS and JEDi do not correspond to qualitative playability, and argue that curating raw gameplay data before training begins and measuring multiple diagnostic properties is a more reliable signal.

Paper proposes Recursive Game Creator, using four recursively iterating components to improve generated game experience

The paper proposes Recursive Game Creator, an experience-oriented agentic game development framework that uses four components — Designer, Builder, Player and Reviewer — to iterate recursively and push a rough game prototype toward a more replayable work.

SpeedrunBench launches to benchmark frontier LLM agents' strategy formation with speedruns across 9 games

The paper introduces SPEEDRUNBENCH, a speedrunning benchmark that evaluates frontier LLM agents on 9 different games. Agents must repeatedly refine their strategies over long-horizon actions, review their own performance and use the knowledge they have gained to be faster than their past selves and others. Experiments show frontier agents approach human world records on simple platformers, but on longer, more complex games they still lag behind human performance under practical budget constraints; the authors see the benchmark as a long-term testbed for studying agents' strategy formation ability.

Attacca: Goal-Directed Control under State Continuity for Long-Horizon Embodied Agents

The researchers propose Attacca, which trains a vision goal-conditioned policy on complete search-to-interaction trajectories so long-horizon embodied agents can carry on to the next task from the position, orientation and world state left by the previous one. The method uses context-decoupled goal sampling, pairing each demonstration with a class-compatible masked goal image from another world, and provides supervision beyond action imitation through a goal mask prediction head, while introducing behavior phase conditioning that distinguishes the Search, Approach and Interact phases.

PlaySuite: a large-scale benchmark for interactive visual intelligence covering more than 5K open-source games

The authors propose PlaySuite, a large-scale benchmark for evaluating interactive visual intelligence built on more than 5K open-source games from PyWeek and itch.io, spanning engines including Pygame, HTML5, Godot and Unity.

Unity demos prototyping a UI Toolkit main menu from concept art with the CLI and an AI agent

Unity has released a tutorial video showing how an AI agent, working through the Unity CLI and starting from concept art and a short brief, can build a usable main menu interface directly inside the Unity editor.

A general option discovery method masters Atari 2600

I'm excited to share Erik's work! "Mastering Atari 2600 Games with Discovered Options": we propose a general option discovery method that, under high-dimensional observations and tabula rasa conditions, learns from a single stream of experience and outperforms DreamerV3, Rainbow and IQN! 1/9 Looks very promising!

Paper proposes MATE, using a joint-attention mechanism to improve multi-agent zero-shot coordination

The paper proposes MATE (Mutual Attention for zero-shot TEaming), a multi-agent reinforcement learning method inspired by human joint attention that lets agents align visual attention on salient objects in the scene during interaction, so they can coordinate zero-shot with unfamiliar partners.

Code2Games: enabling coding agents to generate playable game worlds from natural language

Wei Wu proposes Code2Games, an agent framework that lets coding agents build structured game worlds on top of a Blender world generated from the same game intent. The framework coordinates scene analysis, gameplay planning, constrained game world generation and engine customization through shared scene and gameplay representations, and when adapting to Unreal Engine 5 uses compilation diagnostics, runtime feedback and playtest results for execution-guided refactoring.

Breaking down the AI design in the Command & Conquer and Red Alert source code

AI and Games examined how the AI in Command & Conquer and Red Alert is implemented, drawing on the source code EA released alongside Command & Conquer Remastered Collection and made available on GitHub.

Forza Motorsport's New Drivatar Is Lying to Players

Forza Motorsport's AI was rebuilt from the ground up, and the Drivatar system no longer relies on human play data. In a video, AI and Games argues that what the rebuilt Drivatar shows players is not how it actually works.

Forza's Most Aggressive AI: Drivatar Bowie Knife99

The Drivatar bot Bowie Knife99 in Forza has become a new rival the player community can't stop talking about, known for its aggressive driving style: it rams cars on the road and repeatedly pulls off extreme corner cuts. It is currently the most-discussed AI in the community.

Why Bowie Knife99 Is a Bug in Forza's Drivatar System

In its 85th video, AI and Games explains why Drivatar "Bowie Knife99" in Forza Horizon 6 became both a community meme and a public enemy in online races, tying the phenomenon to Turn 10's changes to AI driver behavior in 2023's Forza Motorsport.