| Game | Mode | Player 1 | Player 2 | Score | Win Rate | Status |
|---|---|---|---|---|---|---|
| Gomoku (五子棋) | AI vs AI | GPT-4o | GPT-4o-mini | 100:0 | 100% | ✅ Tested |
| Gomoku (五子棋) | Human vs AI | Human | GPT-4o | 1:10 | 9.1% | ✅ Tested |
| Word Chain (成语接龙) | AI vs AI | GPT-4o | GPT-4o-mini | -:- | - | 🛑 Testing |
| Word Chain (成语接龙) | Human vs AI | Human | GPT-4o | -:- | - | 🛑 Testing |
| Chinese Chess | AI vs AI | - | - | -:- | - | 🛑 Coming Soon |
| Go | AI vs AI | - | - | -:- | - | 🛑 Coming Soon |
| Chess | AI vs AI | - | - | -:- | - | 🛑 Coming Soon |
Note: Test results are based on multiple game sessions. Individual results may vary.
说明: 测试结果基于多轮游戏对战。单次结果可能有所不同。
An open-source project focused on AI gaming battles using Large Language Models (LLM) technology. Implements intelligent gameplay in classic games with multiple game modes:
- 🤖 AI vs AI: Two different AI models competing against each other
- 👤 Human vs AI: Human players versus AI opponents
- 👥 Multi-model Comparison: Testing strategic capabilities of different AI models
- Multi-model Support: Supports OpenAI GPT-4o, GPT-4o-mini, and other models
- Real-time Battles: Clear console interface with real-time game status
- Intelligent Decision Making: AI analyzes game situations and makes strategic decisions
- Easy Extension: Modular design for easy addition of new games
Location: /gomoku
Classic Gomoku game supporting both AI vs AI and Human vs AI modes.
Game Features:
- Standard 15×15 board
- First to connect five wins
- Maximum 100 moves per player
- Intelligent defense and attack strategies
- Two game modes: AI vs AI and Human vs AI
Location: /word_chain
Intelligent Chinese idiom chain game with strict tail-to-head character matching rules.
Game Features:
- Strict rule: Last character of previous idiom must exactly match first character of next idiom
- Four-character Chinese idioms (成语) only
- Two game modes: AI vs AI and Human vs AI
- Smart validation and scoring system
- Real-time game progress visualization
- Duplicate idiom detection
- Cultural and linguistic challenge
Technical Highlights:
- Clear game state visualization
- Smart idiom validation and character matching
- Detailed game history tracking
- Multi-retry mechanism ensuring effective AI responses
- Chinese character encoding support
Running Example:
=============================================
第 5/100 轮 | 人类: 3 分 | AI: 2 分
当前成语: 龙飞凤舞 | 下一个必须以'舞'开头
=============================================
📖 游戏历史: 一心一意 -> 意气风发 -> 发愤图强 -> 强词夺理 -> 理直气壮
🎯 当前成语: 龙飞凤舞
💡 提示: 下一个成语必须以 '舞' 开头
- Python 3.7+
- OpenAI API Key
-
Clone the Repository
git clone https://github.com/brycewang-stanford/llm-play-games.git cd llm-play-games -
Set up API Key
# Method 1: Environment Variable export OPENAI_API_KEY="your-api-key-here" # Method 2: Input during runtime # The program will prompt for API key input
-
Run Games
Gomoku - AI vs AI Mode:
cd gomoku python3 gomoku.pyGomoku - Human vs AI Mode:
cd gomoku python3 gomoku_human_vs_ai.pyWord Chain - AI vs AI Mode:
cd word_chain python3 word_chain_ai_vs_ai.pyWord Chain - Human vs AI Mode:
cd word_chain python3 word_chain_human_vs_ai.py
- Chinese Chess: Traditional Chinese chess AI battles
- Go: 9×9 or 19×19 Go game battles
- Chess: Classic international chess
- Tic-Tac-Toe: Simple strategy game
- Connect Four: Vertical connection game
- Human vs AI Mode: Support for human player participation ✅ (Available in Gomoku)
- Language: Python 3.7+
- AI Models: OpenAI GPT-4o / GPT-4o-mini
- API: OpenAI API
- Architecture: Modular design
Contributions are welcome! Please follow these steps:
- Fork the Project
- Create your Feature Branch (
git checkout -b feature/AmazingFeature) - Commit your Changes (
git commit -m 'Add some AmazingFeature') - Push to the Branch (
git push origin feature/AmazingFeature) - Open a Pull Request
This project is licensed under the MIT License - see the LICENSE file for details.
这是一个专注于人工智能游戏对战的开源项目。使用大语言模型（LLM）技术，实现经典游戏中的智能对战，支持多种游戏模式：
- 🤖 AI vs AI: 两个不同的AI模型互相对战
- 👤 Human vs AI: 人类玩家与AI对战
- 👥 多模型比较: 测试不同AI模型的策略能力
- 多模型支持: 支持 OpenAI GPT-4o、GPT-4o-mini 等多种模型
- 实时对战: 清晰的控制台界面，实时显示游戏状态
- 智能决策: AI能够分析棋局，做出策略性决策
- 易于扩展: 模块化设计，方便添加新游戏
位置：/gomoku
经典的五子棋游戏，支持AI vs AI和Human vs AI两种模式。
游戏特点：
- 15×15 标准棋盘
- 先连成五子者获胜
- 每位玩家最多100步
- 智能防守与进攻策略
- 两种游戏模式：AI vs AI 和 Human vs AI
位置：/word_chain
智能成语接龙游戏，严格的字字相接规则，具有挑战性。
游戏特点：
- 严格规则：上一个成语的最后一个字必须与下一个成语的第一个字完全相同
- 只能使用四字成语
- 两种游戏模式：AI vs AI 和 Human vs AI
- 智能验证和计分系统
- 实时游戏进度可视化
- 重复成语检测
- 文化和语言挑战
技术亮点：
- 清晰的游戏状态可视化显示
- 智能的成语验证与字符匹配
- 详细的游戏状态追踪
- 多轮重试机制确保AI响应有效
- 中文字符编码支持
运行示例：
=============================================
第 5/100 轮 | 人类: 3 分 | AI: 2 分
当前成语: 龙飞凤舞 | 下一个必须以'舞'开头
=============================================
📖 游戏历史: 一心一意 -> 意气风发 -> 发愤图强 -> 强词夺理 -> 理直气壮
🎯 当前成语: 龙飞凤舞
💡 提示: 下一个成语必须以 '舞' 开头
- Python 3.7+
- OpenAI API 密钥
-
克隆项目
git clone https://github.com/brycewang-stanford/llm-play-games.git cd llm-play-games -
设置API密钥
# 方法1：环境变量 export OPENAI_API_KEY="your-api-key-here" # 方法2：程序运行时输入 # 运行程序时会提示输入API密钥
-
运行游戏
五子棋 - AI vs AI 模式:
cd gomoku python3 gomoku.py五子棋 - 人机对战模式:
cd gomoku python3 gomoku_human_vs_ai.py成语接龙 - AI vs AI 模式:
cd word_chain python3 word_chain_ai_vs_ai.py成语接龙 - 人机对战模式:
cd word_chain python3 word_chain_human_vs_ai.py
- 象棋 (Chinese Chess): 传统中国象棋AI对战
- 围棋 (Go): 9×9或19×19围棋对战
- 国际象棋 (Chess): 经典国际象棋
- 井字棋 (Tic-Tac-Toe): 简单策略游戏
- 四子棋 (Connect Four): 垂直连线游戏
- 人机对战模式: 支持人类玩家参与 ✅ (五子棋已实现)
- 语言: Python 3.7+
- AI模型: OpenAI GPT-4o / GPT-4o-mini
- API: OpenAI API
- 架构: 模块化设计
欢迎贡献代码！请遵循以下步骤：
- Fork 本项目
- 创建特性分支 (
git checkout -b feature/AmazingFeature) - 提交更改 (
git commit -m 'Add some AmazingFeature') - 推送到分支 (
git push origin feature/AmazingFeature) - 创建 Pull Request
本项目采用 MIT 许可证 - 查看 LICENSE 文件了解详情。
Project Link: https://github.com/brycewang-stanford/llm-play-games
⭐ If this project helps you, please give it a star! ⭐ 如果这个项目对你有帮助，请给它一个星标！
Note: The Splendor game in this repository is incomplete as the author is still clarifying the official rules.