Skip to content

Fields

Game agents latest news

AI that acts in games: generalist game agents, LLM-driven bots, agents in Minecraft, StarCraft and other environments, benchmarks and open frameworks.

111 picks55 in the last 30 days170 collected in all

Updated

Picks archive · Page 4

Oct 22Thu61–80
  1. arXiv Board & Classic Game AI39

    ToM2Vec: Continuous theory-of-mind tokens for multi-agent coordination

    ToM2Vec learns theory of mind for multi-agent coordination through a fixed budget of continuous reasoning tokens, using partners' actual thoughts as training supervision. Across cooperative, competitive, and mixed-motive environments, including Overcooked-AI, multi-party negotiation, Decrypto, Hanabi, and Collab Escape/Capture, it delivers consistent gains over the base model and several competitive baselines.

    Why it matters: The article compresses theory of mind into a fixed number of continuous reasoning tokens, supervised by partners' actual thoughts, in contrast to text-based reasoning approaches.

Oct 21Wed
  1. 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.

Oct 20Tue
  1. arXiv Game AI58

    Scaling in games: continued pre-training for embodied agents in diverse virtual worlds

    The paper systematically studies how data volume, diversity, and model scale affect embodied agents in continued pre-training, using 120 training configurations and models from 0.8B to 27B parameters. Diversity improves transfer to related worlds but gives limited gains for worlds that differ substantially. Larger models improve data efficiency, but when trained on a single world they generalize worse to distant worlds. The authors argue that building general-purpose embodied agents requires training stages richer than continued pre-training.

    Why it matters: The paper compares the effects of data volume, diversity, and model scale across 120 training configurations. Its finding that gains from cross-world transfer are limited makes it a useful reference for game agent training strategies.

  2. arXiv Game AI36

    CHSM8: a chess Transformer that follows natural-language style prompts

    CHSM8 is a 210M-parameter chess Transformer that selects moves based on free-text style descriptions such as "Sicilian Defense" or "solid and conservative." Without search, it reaches 1837 Elo against a strength-limited Stockfish. Supervised fine-tuning and DPO improved its adherence to style; for example, the early queen-trade rate rose from 53% to 67%. The paper also releases the weights, a training-labeled dataset, prompts, configuration, and inference code.

    Why it matters: The paper releases its weights, datasets, prompts, and code, and documents failed ablation experiments, making its training pipeline easier to reproduce and learn from.

  3. arXiv Game AI39

    KPHarness: Co-Evolving Knowledge and Probing for Long-Horizon Gaming Harnesses

    The paper proposes KPHarness, which co-evolves knowledge and probing to rewrite the external execution harness of a frozen VLA, so that it not only swaps intervention rules but also improves how it investigates failures. On mining, combat, and crafting tasks in Minecraft, it achieves broader task coverage than pure policy execution, a hand-specified fixed harness, and a version with knowledge evolution only and no adaptive probing; each harness revision is evaluated pairwise against the existing version under matched sandbox conditions and is accepted only if it shows improvement.

    Why it matters: The paper presents a way to feed failure-diagnosis evidence into harness revision, and compares it against a fixed harness and a knowledge-evolution-only baseline.

  4. YYS (Youyanshe)47

    20 large models take on Slay the Spire at A20 with only one clear, losses mostly blamed on being too cautious

    Researchers used SpireBench to have 20 large models play through the same 10 new seeds at A20 difficulty in Slay the Spire; across 200 runs, only GPT-5.6 Sol cleared the game once. Compared with four top streamers, the AI fought only about 0.9-1.4 elite battles in Act 1 on average, versus about 2.6 for the streamers, and the AI chose to heal at 57%-75% of campfires, versus about 9% for the experts. Counting the first act alone, the best-performing models are already close to top human players, but long-term strategic ability still shows a clear gap.

    Why it matters: The article traces the models' failures to excessive caution and sets them against per-run data from top streamers, making the models' specific weaknesses in long-term resource management easy to grasp.

Oct 19Mon
  1. arXiv Board & Classic Game AI34

    Researchers propose SPOKE, training poker agents with strategic preference optimization to reach the lowest average exploitability on Kuhn and Leduc

    The researchers propose SPOKE (Strategic Preference Optimization with KL-anchored Equilibrium). Across three runs of the SmallThinker-4B-A0.6B policy, it achieved the lowest average exploitability of 0.213 on Kuhn and 1.081 on Leduc, with a six-player table top-1 agreement rate of 0.775.

    Why it matters: The paper compares SPOKE against SFT, DPO, GRPO and GSPO and reports average exploitability figures on Kuhn and Leduc, letting readers judge how the method differs from other policy optimization approaches.

  2. Naavik61

    Meta's role in gaming today: Muse, VR, and UA

    Meta is becoming an important user acquisition channel for games again, with its share of social media game ad impressions rising from 54% to 69% in 2025. Meanwhile, its VR gaming and Horizon UGC businesses have been clearly deprioritized, and the VR Glasses launch mentioned games in just one sentence. The article also argues that personal agents like Muse could change how players discover games and make in-game purchases, but their direct connection to games is still unclear.

    Why it matters: This article looks at Meta's resurgence in user acquisition alongside the shrinking of its VR business, offering a clear clue for judging where game user acquisition channels are heading.

Oct 18Sun
  1. arXiv Game AI58

    CHSM8: a chess Transformer that follows natural-language style prompts

    CHSM8 is a 210M-parameter chess Transformer that selects moves based on free-text style descriptions such as "Sicilian Defense" or "solid and conservative." Without search, it reaches 1837 Elo against a strength-limited Stockfish. Supervised fine-tuning and DPO improved its adherence to style; for example, the early queen-trade rate rose from 53% to 67%. The paper also releases the weights, a training-labeled dataset, prompts, configuration, and inference code.

    Why it matters: The paper releases its weights, datasets, prompts, and code, and documents failed ablation experiments, making its training pipeline easier to reproduce and learn from.

Oct 17Sat
  1. arXiv Game AI43

    Mirror: self-improving video-to-action models

    Mirror is a self-supervised method that learns video-to-action mapping through interaction with the environment, without prior action data or hand-written rewards. On Atari, OGBench, and DeepMind Control, it outperforms mainstream observation-based learning methods using its learned video plans. On Minecraft BASALT, it surpasses methods that have access to expert actions or hand-crafted task knowledge. In Minecraft survival tasks, it obtained an iron pickaxe through self-improvement alone, with no reward or expert action labels.

    Why it matters: It doesn't rely on human action annotations or hand-written rewards, letting video planning models become executable agent policies. Worth a look for readers interested in combining video models with agents.

  2. arXiv Game AI45

    Mirror: self-improving video-to-action models

    Mirror is a self-supervised method that learns video-to-action mapping through interaction with the environment, without prior action data or hand-written rewards. On Atari, OGBench, and DeepMind Control, it outperforms mainstream observation-based learning methods using its learned video plans. On Minecraft BASALT, it surpasses methods that have access to expert actions or hand-crafted task knowledge. In Minecraft survival tasks, it obtained an iron pickaxe through self-improvement alone, with no reward or expert action labels.

    Why it matters: It doesn't rely on human action annotations or hand-written rewards, letting video planning models become executable agent policies. Worth a look for readers interested in combining video models with agents.

Oct 16Fri
  1. arXiv Board & Classic Game AI37

    Explaining chess and Go in natural language, with the BinProp method

    The paper proposes an automatic metric based on chess engine evaluation and finds that current mainstream large language models perform markedly poorly at explaining chess moves, often relying on surface features of the board. The proposed BinProp is a training-free method that lets such models discover reusable chess concepts; its average score is nearly triple that of direct prompting. The same method also discovers useful concepts in Go, checkers, and shogi, without additional tuning.

    Why it matters: The method needs no additional training yet lets a language model discover reusable chess concepts, offering a useful reference for approaches to explaining engine moves.

  2. arXiv Game AI34

    Paper builds controlled Atari and sente evaluation tasks to test the language-based Go model LoGos

    The researchers built a controlled evaluation set around the relationship between atari and sente in Go, to test whether language models can recognize these terms and connect them to move evaluation. The subject of the test is LoGos, a Go language model supervised-fine-tuned on terminology and commentary data, and the results show it struggles both to recognize changes in atari relationships and to judge the strategic meaning of such changes.

    Why it matters: The authors use controlled local board modifications to separate term recognition from move evaluation, offering a reusable test approach for checking whether language models can ground domain terminology in a specific position.

  3. arXiv Game AI63

    A team of Gemini 4 Argon agents beats Minecraft

    Four Gemini 4 Argon agents collaborated autonomously in a Minecraft survival world, defeating the Ender Dragon in 27 hours and 14 minutes. The team wrote 2,783 programs in a random world on easy difficulty and died 605 times along the way; researchers could watch in real time and pause at any point, but no intervention was needed. To the authors' knowledge, this is the first time AI agents have completed Minecraft end to end autonomously, using only the standard Mineflayer library and no hand-picked map seed or game-specific tools.

    Why it matters: The paper documents how four agents completed Minecraft with no human-designed game logic, and their division of labor, experience sharing and long-horizon planning are worth studying.

  4. arXiv Game AI45

    Chloe Su proposes MFM multimodal filesystem memory architecture, outperforming baselines on long-horizon embodied agent tasks

    The paper proposes MFM, a multimodal filesystem memory architecture that stores episodic, spatial, semantic and procedural, and working memory separately, and links these representations to the underlying multimodal segments through retrievable text records, so that an agent can use contextual cues to find past experiences and revisit the original observations to recover details missing from the text descriptions.

    Why it matters: The paper lays out the design of layered memory plus text retrieval linked back to the original segments, and compares three base models on Minecraft and computer-operation tasks.

  5. arXiv Game AI73

    A team of Gemini 4 Argon agents beats Minecraft

    Four Gemini 4 Argon agents collaborated autonomously in a Minecraft survival world, defeating the Ender Dragon in 27 hours and 14 minutes. The team wrote 2,783 programs in a random world on easy difficulty and died 605 times along the way; researchers could watch in real time and pause at any point, but no intervention was needed. To the authors' knowledge, this is the first time AI agents have completed Minecraft end to end autonomously, using only the standard Mineflayer library and no hand-picked map seed or game-specific tools.

    Why it matters: The paper documents how four agents completed Minecraft with no human-designed game logic, and their division of labor, experience sharing and long-horizon planning are worth studying.

Oct 15Thu
  1. arXiv Game AI41

    Player-1: Learning to play from synthetic experience

    Player-1 is a post-training framework that uses synthetic experience to let game agents learn game operations beyond the scope of their demonstration data. On MineDojo-300, Player-1-9B and Player-1-35B-A3B improve over their respective SFT policies by 27.8 and 17.7 percentage points. Player-1-9B reaches a 38.33% success rate, while the strongest baseline reaches 12.63%.

    Why it matters: This paper expands game agent training from large amounts of demonstration data to reinforcement learning across 16 synthetic environments, offering a useful contrast for thinking about moving beyond reliance on demonstrations.

  2. arXiv Board & Classic Game AI35

    Why accuracy fails to certify theory of mind in multi-agent teams

    The paper argues that modules in multi-agent cooperation that infer actions from teammates should not be certified on accuracy alone, because the cost of errors and the correlation between errors change the acceptable accuracy threshold. The authors tested this in task allocation, Level-Based Foraging and Hanabi, finding that accuracy fails in both directions. In Hanabi, where errors compound, it approved a harmful LLM deployment.

    Why it matters: The paper shifts the certification standard for theory-of-mind modules from predictive accuracy to deployment cost and error dependence, offering a direct reference for evaluating whether collaborative AI can be released.

Oct 14Wed
  1. arXiv Game World Models & Generation35

    Paper proposes CGWMs, learning interpretable world model concept representations with temporal logic

    The paper proposes Concept-Grounded World Models (CGWMs), which learn interpretable concept representations from observations using complete knowledge expressed in temporal logic rather than black-box representations. Planning is formulated as differentiable inference, optimizing agent actions via gradient ascent to satisfy goals. In comparisons with pure-neural and neurosymbolic baselines in two game environments, CGWMs' concept grounding is more accurate, holds over long horizons as well, and plans more robustly both in and out of distribution.

    Why it matters: The paper places temporal-logic knowledge in the same framework as concept learning and planning for world models, and compares against pure-neural and neurosymbolic baselines in two game environments, making it a basis for judging the method trade-offs of this line of work.

  2. PocketGamer.biz49

    Playgama launches tools for AI agents to publish and improve games

    Playgama has launched a set of tools that let AI agents publish games, analyze player behavior and submit updated versions. Its MCP handles game uploads and returns a shareable online link, while Playgama Sandbox provides discovery, data analytics, live ad revenue and paid user acquisition. The updated Bridge SDK lets agents use player data to guide later improvements, and the company plans to add more tools and services in the coming months.

    Why it matters: The report connects AI agent publishing, player data feedback, and platform distribution into one pipeline, which can be read alongside similar platforms' capabilities for comparison.