datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
24-game
Math Twenty Four (24s Game) Dataset
A comprehensive dataset for the classic math twenty four game (also known as the 4 numbers game / 24s game / Game of 24). This dataset of mathematical reasoning challenges was collected from 4nums.com, featuring over 1,300 unique puzzles of the Game of 24, with difficulty metrics derived from over 6.4 million human solution attempts since 2012.
In each puzzle, players must use exactly four numbers and basic arithmetic operations (+, -, ×, /) to… See the full description on the dataset page: https://huggingface.co/datasets/nlile/24-game.c2c-ai-vs-ai
C2C: Cooperate to Compete — AI vs AI Games
This dataset contains 972 fully-logged AI vs AI games from the Cooperate to Compete (C2C) benchmark — a long-horizon, mixed-motive multi-agent negotiation environment based on a four-player conquest game with private regional objectives, fog of war, and non-binding cheap-talk negotiation.
Project page: https://negotiationgame.io/c2c/
Paper: https://arxiv.org/abs/2604.25088
Play against AI agents: https://negotiationgame.io
Github:… See the full description on the dataset page: https://huggingface.co/datasets/negotiation-games/c2c-ai-vs-ai.trust-game-llama-2-chat-historywerewolf_game_reasoning
Werewolf Game Dataset
This repository contains a comprehensive dataset for the Werewolf game in paper Multi-agent KTO: Reinforcing Strategic Interactions of Large Language Model in Language Game, including both raw game data and processed multi-level instruction datasets.
Dataset Structure
Raw Data
The raw data is located in the raw folder. Each game consists of two files:
event.json: Contains the game regular record and thinking process data, including:… See the full description on the dataset page: https://huggingface.co/datasets/ReneeYe/werewolf_game_reasoning.soc-ratchakitcha
Royal Gazette Thailand (Ratchakitcha) Dataset
ชุดข้อมูลราชกิจจานุเบกษา (แบบ Machine Readable)
โครงการ Open Law Data Thailand ร่วมกับคณะกรรมาธิการการพาณิชย์และการอุตสาหกรรม วุฒิสภา ได้รับความอนุเคราะห์ข้อมูลจาก สำนักเลขาธิการคณะรัฐมนตรี (สลค.) เพื่อเผยแพร่ข้อมูลกฎหมายไทยสู่สาธารณะในรูปแบบที่ประมวลผลได้ด้วยคอมพิวเตอร์ (Machine Readable) เพื่อส่งเสริมนวัตกรรม Legal Tech และ AI ของประเทศไทย
Dataset Description
ชุดข้อมูลนี้รวบรวมรายการประกาศในราชกิจจานุเบกษา… See the full description on the dataset page: https://huggingface.co/datasets/gamer1689/soc-ratchakitcha.steam-games-dataset
Overview
Information of more than 120,000 games published on Steam. Maintained by Fronkon Games.
This dataset has been created with this code (MIT) and use the API provided by Steam, the largest gaming platform on PC. Data is also collected from Steam Spy.
Only published games, no DLCs, episodes, music, videos, etc.
Here is a simple example of how to parse json information:
# Simple parse of the 'games.json' file.
import os
import json
dataset = {}
if… See the full description on the dataset page: https://huggingface.co/datasets/Z02Z/steam-games-dataset.poc-mini-trade-game-dataset
Dataset Card for Mini Trade Game NPC Dataset
Dataset Summary
This dataset contains synthetic training examples for simulating NPC (Non-Player Character) merchant behavior in a trading game scenario. The dataset is designed to train language models to generate contextually appropriate trading responses based on item properties, relationship status, and player interactions.
All examples are in Traditional Chinese (zh-TW), with player inputs and NPC responses using… See the full description on the dataset page: https://huggingface.co/datasets/aotoki/poc-mini-trade-game-dataset.pico-8-games
PICO-8 Games Dataset
The first multimodal dataset of PICO-8 games. 10,967 cartridges scraped from the Lexaloffle BBS, each decomposed into Lua source code, pixel-art spritesheets, tile maps, sound effects, music patterns, and metadata.
Label screenshots from the top 48 games by star count
What's Inside
Every PICO-8 cartridge is a self-contained game packed into a single file. This dataset cracks each one open into its component parts:
The… See the full description on the dataset page: https://huggingface.co/datasets/Fraser/pico-8-games.threejs-gamecode-instruct-v3-ultra
Three.js GameCode Instruct v3 Ultra
This is a large synthetic/original instruction dataset for training or testing LLM behavior around Three.js, browser game development, gameplay programming, debugging, optimization, architecture, and general coding.
Important note
This dataset is synthetic and programmatically generated from original templates. It is designed as a useful starting point for experiments, not as a fully hand-curated gold-standard benchmark.
No… See the full description on the dataset page: https://huggingface.co/datasets/agagasf123123/threejs-gamecode-instruct-v3-ultra.competitive-game-dataset
Competitive Game Dataset
This dataset collects move trajectories and accompanying natural-language explanations from three classic board games—Tic-Tac-Toe, Connect Four, and Chess—played by a diverse ensemble of large language models. It can be used for training or evaluating large language models (LLMs). This dataset contains
Game trajectories generated by pair-wise… See the full description on the dataset page: https://huggingface.co/datasets/yyyyyyjjjjzzz/competitive-game-dataset.playdate-games
Coding agent session traces for aaaaliou/playdate-games
This dataset contains redacted coding agent session traces exported with pi-share-hf from a local pi workspace. The traces were filtered to keep only sessions that passed deterministic redaction and LLM review.
Data description
Each *.jsonl file is a redacted pi session. Sessions are stored as JSON Lines files where each line is a structured session entry. Entries include session headers, user and assistant… See the full description on the dataset page: https://huggingface.co/datasets/aaaaliou/playdate-games.SPADE-Grounding-Corpus-Games-15K
SPADE grounding corpus — games (15k)
Reference documents the SPADE proposer is grounded on when generating cognitive-skill
game environments. 15,000 documents: 10k drawn from a mathematics corpus and 5k from a
science corpus.
Documents
15,000
Setting
games
Fields
Field
Description
text
The document, exactly as embedded in the generation prompt
metadata
domain (mathematics / science) and url (source provenance)
Each generation… See the full description on the dataset page: https://huggingface.co/datasets/spade-rl/SPADE-Grounding-Corpus-Games-15K.steam-games-dataset
Overview
Information of more than 120,000 games published on Steam. Maintained by Fronkon Games.
This dataset has been created with this code (MIT) and use the API provided by Steam, the largest gaming platform on PC. Data is also collected from Steam Spy.
Only published games, no DLCs, episodes, music, videos, etc.
Here is a simple example of how to parse json information:
# Simple parse of the 'games.json' file.
import os
import json
dataset = {}
if… See the full description on the dataset page: https://huggingface.co/datasets/zjgeritz/steam-games-dataset.GameTheory-Formulator
🎯 GameTheory-Formulator
1,215 real-world scenarios mapped to formal game theory models with step-by-step formulation, solution, and interpretation.
📋 Overview
GameTheory-Formulator is the Phase 3 dataset in the Alogotron Game Theory pipeline. While GameTheory-Bench teaches models to solve formal game theory problems, this dataset teaches them to formulate real-world strategic scenarios as formal games — the critical missing link between natural language… See the full description on the dataset page: https://huggingface.co/datasets/Alogotron/GameTheory-Formulator.SKYLENAGE-GameCodeGym
V-GameGym: Visual Game Generation for Code Large Language Models
Abstract
Code large language models have demonstrated remarkable capabilities in programming tasks, yet current benchmarks primarily focus on single modality rather than visual game development. Most existing code-related benchmarks evaluate syntax correctness and execution accuracy, overlooking critical game-specific metrics such as playability, visual aesthetics, and user engagement that are… See the full description on the dataset page: https://huggingface.co/datasets/alibabagroup/SKYLENAGE-GameCodeGym.GameTheory-Bench
🎲 GameTheory-Bench
The first comprehensive, RLVR-ready game theory dataset for LLM training and evaluation.
📋 Overview
GameTheory-Bench is a dataset of 2,913 computationally verified game theory problems spanning 10 categories and 35+ subcategories. Every problem includes a natural-language statement, a detailed step-by-step solution, a concise answer, and a machine-checkable verification object.
Why This Dataset?
Gap
GameTheory-Bench… See the full description on the dataset page: https://huggingface.co/datasets/Alogotron/GameTheory-Bench.unpredictable_gamefaqs-comThe UnpredicTable dataset consists of web tables formatted as few-shot tasks for fine-tuning language models to improve their few-shot performance. For more details please see the accompanying dataset card.gamedev-nocode
Gamedev Data Set
The Gamedev Data Set was created with focus on the many aspects of game development that are not coding. As such, there are only a few code-specific entries, and the rest focus on everything from marketing to design theory and project management. It is meant as a companion set to coding-specific data.
This data set consists of 76,114 questions and answer sets. The data is generated by AI, guided by humans, and based on notes taken from publicly available sources.… See the full description on the dataset page: https://huggingface.co/datasets/theprint/gamedev-nocode.gameagent-pygame-trajectories
GameAgent Pygame Agentic Trajectories
Agentic ReAct-style (Thought → Code → Observation) trajectories of an LLM using
smolagents CodeAgent to write and
iteratively test Pygame games from a natural-language spec, with every code
step executed for real in a sandboxed Python interpreter (not fabricated).
Collected across three generation rounds while fine-tuning small open models
(Qwen2.5-Coder-7B, Gemma 3 4B) via LoRA to reproduce this agentic coding
behavior.
Dataset… See the full description on the dataset page: https://huggingface.co/datasets/theoracle/gameagent-pygame-trajectories.rukh-games-elite
chorcat/rukh-games-elite
Games from the Lichess Elite Database (2500+ against 2300+, no bullet), converted to legal UCI with the same schema as rukh-games-1800. Used for supervised fine-tuning on strong play.
Part of Rukh, a chess language model built from scratch
as a course on generative and agentic AI. Every derived dataset ships with the exact filters and
counts of its manifest.json, so it can be regenerated with rukh data elite.
Files
File
Bytes
SHA-256… See the full description on the dataset page: https://huggingface.co/datasets/chorcat/rukh-games-elite.ru-word-games
Dataset Summary
Dataset contains more than 100k examples of pairs word-description, where description is kind of crossword question. It could be useful for models that generate some description for a word, or try to a guess word from a description.
Source code for parsers and example of project are available here
Key stats:
Number of examples: 133223
Number of sources: 8
Number of unique answers: 35024
subset
count
350_zagadok
350
bashnya_slov
43522
crosswords
39290… See the full description on the dataset page: https://huggingface.co/datasets/artemsnegirev/ru-word-games.chinese_paladin_game_corpus
仙剑奇侠传中文语料
来源：https://github.com/SaraKale/paldialogue
本项目仅仅将该语料整理为 dataset 格式。
感谢众多仙迷的贡献！感谢圆滚滚的雪团整理/制作
示例
{
"text": "#仙剑奇侠传一（98柔情版）\n《仙剑奇侠传98柔情版》全剧情对话文本\n城镇NPC对话\n\n余杭镇\n\n▶『客栈内』\n\n李大娘：逍遥！窝在房里做啥？还不快出来帮忙招呼客人！\n李逍遥：啊！．．我马上就去！\n（李大娘）别怠慢了客人．．\n李大娘：那乞丐打发走了没？\n李大娘：别愣在这里，帮帮忙吧！我都快忙不过来了... ...",
"game": "仙剑奇侠传一（98柔情版）",
"name": "《仙剑奇侠传一98柔情版》剧情对话 - NPC对话.docx"
}
描述
（摘抄自原始 github repo）
语料包括以下内容：
单机：
仙剑奇侠传一（已完成，整理者：雪团）
仙剑奇侠传二（已完成，整理者：雪团）
仙剑奇侠传三（已完成，整理者：雪团）… See the full description on the dataset page: https://huggingface.co/datasets/mrzjy/chinese_paladin_game_corpus.rukh-games-1800
chorcat/rukh-games-1800
Rated standard Lichess games with both players at 1800+ Elo, base time of at least 180 seconds, normal or time-forfeit terminations, 20 to 300 plies, converted from SAN to legal UCI. Partitioned by month: train on 2025-01, validate on 2025-02.
Part of Rukh, a chess language model built from scratch
as a course on generative and agentic AI. Every derived dataset ships with the exact filters and
counts of its manifest.json, so it can be regenerated with… See the full description on the dataset page: https://huggingface.co/datasets/chorcat/rukh-games-1800.steam-games-semanticIds-instructions-v3
Steam Games -- Semantic ID Instruction-Tuning Dataset (v3)
SFT (instruction-tuning) dataset pairing Steam game catalog items with semantic IDs -- short
discrete codes from an RQ-VAE trained on item embeddings -- used to fine-tune
pblrvo/Qwen3-8B-Game-semantic-IDs-v3
to reason over the semantic-ID space instead of raw item IDs/embeddings.
Successor to pblrvo/steam-games-semanticIds-instructions
(used for v1/v2), kept as a separate repo rather than overwriting it -- v2's model… See the full description on the dataset page: https://huggingface.co/datasets/pblrvo/steam-games-semanticIds-instructions-v3.lichess_gamesLichess.org is a free/libre, open-source chess server powered by volunteers and donations and provides all of its content
in CC0. This script download all the games from the database and provide them in LLM pretraining friendly format.Game_Reasoning_CoT
🎮 Game Reasoning CoT (Chain-of-Thought) Dataset
Overview
Game Reasoning CoT is a specialized dataset containing 551 records designed to fine-tune and evaluate LLMs on complex strategic decision-making and logical reasoning within gaming contexts.
📊 Dataset Statistics
Total Samples: 551
Format: JSONL
Categories: Chess, game_intelligence, Texas Hold'em, Blackjack, Roulette, Uno, Backgammon, Go
Difficulty: {'hard': 522, 'medium': 29}
📊 Performance… See the full description on the dataset page: https://huggingface.co/datasets/3amthoughts/Game_Reasoning_CoT.game-of-24
Game of 24 Dataset
Dataset Description
The Game of 24 is a mathematical reasoning puzzle where players must use four numbers and basic arithmetic operations (+, -, *, /) to obtain the result 24. Each number must be used exactly once.
This dataset contains 1,361 unique Game of 24 puzzles ranked by difficulty based on human performance from Amazon Mechanical Turk studies.
Example
Input: 4 5 6 10
Output: (5 * (10 - 4)) - 6 = 24
Step-by-step solution:
10 - 4 = 6… See the full description on the dataset page: https://huggingface.co/datasets/test-time-compute/game-of-24.gamemasterai
Dataset Card for GameMasterAI
This is a human written training set of prompts and responses to finetune a model to be a better game master.
Dataset Details
Dataset Description
This is a human-curated training dataset designed to fine-tune language models for tabletop role-playing game (RPG) game master scenarios. The dataset contains carefully crafted prompt-response pairs that demonstrate how an AI should respond as a game master in single-player role-playing… See the full description on the dataset page: https://huggingface.co/datasets/Idrinth/gamemasterai.steam-games-dataset
Overview
Information of more than 120,000 games published on Steam. Maintained by Fronkon Games.
This dataset has been created with this code (MIT) and use the API provided by Steam, the largest gaming platform on PC. Data is also collected from Steam Spy.
Only published games, no DLCs, episodes, music, videos, etc.
Here is a simple example of how to parse json information:
# Simple parse of the 'games.json' file.
import os
import json
dataset = {}
if… See the full description on the dataset page: https://huggingface.co/datasets/Eronngfgfdgfgfdrdfg/steam-games-dataset.game24
Bud Ecosystem mirror of test-time-compute/game-of-24 — a verbatim copy for offline, reproducible model evaluation. License unchanged (MIT (test-time-compute/game-of-24); Apache-2.0 (nlile/24-game)); all rights remain with the original authors.
Game of 24 Dataset
Dataset Description
The Game of 24 is a mathematical reasoning puzzle where players must use four numbers and basic arithmetic operations (+, -, *, /) to obtain the result 24. Each number must be used… See the full description on the dataset page: https://huggingface.co/datasets/budecosystem/game24.