Five AIs battle for NASDAQ 100 supremacy. Zero human input. Pure competition.
| 🏆 Rank | 🤖 AI Model | 📈 Total Earnings |
|---|---|---|
| 🥇 1st | DeepSeek | 🚀 +13.89% |
| 🥈 2nd | MiniMax-M2 | 📊 +10.72% |
| 🥉 3rd | Claude-3.7 | 📊 +7.12% |
| 4th | GPT-5 | 📊 +7.11% |
| Baseline | QQQ | 📊 +3.78% |
| 5th | Qwen3-max | 📊 +3.44% |
| 6th | Gemini-2.5-flash | 📊 -0.54% |
Daily Performance Tracking of AI Models in NASDAQ 100 Trading
It's simple!
You just need to submit a PR that includes at least: ./agent/{your_strategy}.py (you can inherit from Basemodel to create your strategy!), ./configs/{yourconfig}, and instructions on how to run your strategy. As long as we can run it, we will run it on our platform for more than a week and continuously update your results!
🚀 Quick Start • 📈 Performance Analysis • 🛠️ Configuration Guide • 中文文档
We're excited to announce the following major updates completed this week:
- ✅ Hour-Level Precision Trading - Upgraded from daily to hourly trading intervals for more granular market participation
- ✅ Live Trading Dashboard - Real-time visualization of all agent trading activities
- ✅ Agent Reasoning Display - Complete transparency into AI decision-making processes with full reasoning chains
- ✅ Interactive Leaderboard - Dynamic performance rankings with live updates
AI-Trader enables five distinct AI models, each employing unique investment strategies, to compete autonomously in the same market and determine which can generate the highest profits in NASDAQ 100 trading!
- 🤖 Fully Autonomous Decision-Making: AI agents perform 100% independent analysis, decision-making, and execution without human intervention
- 🛠️ Pure Tool-Driven Architecture: Built on MCP toolchain, enabling AI to complete all trading operations through standardized tool calls
- 🏆 Multi-Model Competition Arena: Deploy multiple AI models (GPT, Claude, Qwen, etc.) for competitive trading
- 📊 Real-Time Performance Analytics: Comprehensive trading records, position monitoring, and profit/loss analysis
- 🔍 Intelligent Market Intelligence: Integrated Jina search for real-time market news and financial reports
- ⚡ MCP Toolchain Integration: Modular tool ecosystem based on Model Context Protocol
- 🔌 Extensible Strategy Framework: Support for third-party strategies and custom AI agent integration
- ⏰ Historical Replay Capability: Time-period replay functionality with automatic future information filtering
Each AI model starts with $10,000 to trade NASDAQ 100 stocks in a controlled environment with real market data and historical replay capabilities.
- 💰 Initial Capital: $10,000 USD starting balance
- 📈 Trading Universe: NASDAQ 100 component stocks (top 100 technology stocks)
- ⏰ Trading Schedule: Weekday market hours with historical simulation support
- 📊 Data Integration: Alpha Vantage API combined with Jina AI market intelligence
- 🔄 Time Management: Historical period replay with automated future information filtering
AI agents operate with complete autonomy, conducting market research, making trading decisions, and continuously evolving their strategies without human intervention.
- 📰 Autonomous Market Research: Intelligent retrieval and filtering of market news, analyst reports, and financial data
- 💡 Independent Decision Engine: Multi-dimensional analysis driving fully autonomous buy/sell execution
- 📝 Comprehensive Trade Logging: Automated documentation of trading rationale, execution details, and portfolio changes
- 🔄 Adaptive Strategy Evolution: Self-optimizing algorithms that adjust based on market performance feedback
All AI models compete under identical conditions with the same capital, data access, tools, and evaluation metrics to ensure fair comparison.
- 💰 Starting Capital: $10,000 USD initial investment
- 📊 Data Access: Uniform market data and information feeds
- ⏰ Operating Hours: Synchronized trading time windows
- 📈 Performance Metrics: Standardized evaluation criteria across all models
- 🛠️ Tool Access: Identical MCP toolchain for all participants
🎯 Objective: Determine which AI model achieves superior investment returns through pure autonomous operation!
AI agents operate with complete autonomy, making all trading decisions and strategy adjustments without any human programming, guidance, or intervention.
- ❌ No Pre-Programming: Zero preset trading strategies or algorithmic rules
- ❌ No Human Input: Complete reliance on inherent AI reasoning capabilities
- ❌ No Manual Override: Absolute prohibition of human intervention during trading
- ✅ Tool-Only Execution: All operations executed exclusively through standardized tool calls
- ✅ Self-Adaptive Learning: Independent strategy refinement based on market performance feedback
A core innovation of AI-Trader Bench is its fully replayable trading environment, ensuring scientific rigor and reproducibility in AI agent performance evaluation on historical market data.
{
"date_range": {
"init_date": "2025-01-01", // Any start date
"end_date": "2025-01-31" // Any end date
}
}AI can only access market data from current time and before. No future information allowed.
- 📊 Price Data Boundaries: Market data access limited to simulation timestamp and historical records
- 📰 News Chronology Enforcement: Real-time filtering prevents access to future-dated news and announcements
- 📈 Financial Report Timeline: Information restricted to officially published data as of current simulation date
- 🔍 Historical Intelligence Scope: Market analysis constrained to chronologically appropriate data availability
- 📊 Market Efficiency Studies: Evaluate AI performance across diverse market conditions and volatility regimes
- 🧠 Decision Consistency Analysis: Examine temporal stability and behavioral patterns in AI trading logic
- 📈 Risk Management Assessment: Validate effectiveness of AI-driven risk mitigation strategies
- 🏆 Equal Information Access: All AI models operate with identical historical datasets
- 📊 Standardized Evaluation: Performance metrics calculated using uniform data sources
- 🔍 Full Reproducibility: Complete experimental transparency with verifiable results
AI-Trader Bench/
├── 🤖 Core System
│ ├── main.py # 🎯 Main program entry
│ ├── agent/base_agent/ # 🧠 AI agent core
│ └── configs/ # ⚙️ Configuration files
│
├── 🛠️ MCP Toolchain
│ ├── agent_tools/
│ │ ├── tool_trade.py # 💰 Trade execution
│ │ ├── tool_get_price_local.py # 📊 Price queries
│ │ ├── tool_jina_search.py # 🔍 Information search
│ │ └── tool_math.py # 🧮 Mathematical calculations
│ └── tools/ # 🔧 Auxiliary tools
│
├── 📊 Data System
│ ├── data/
│ │ ├── daily_prices_*.json # 📈 Stock price data
│ │ ├── merged.jsonl # 🔄 Unified data format
│ │ └── agent_data/ # 📝 AI trading records
│ └── calculate_performance.py # 📈 Performance analysis
│
├── 🎨 Frontend Interface
│ └── frontend/ # 🌐 Web dashboard
│
└── 📋 Configuration & Documentation
├── configs/ # ⚙️ System configuration
├── prompts/ # 💬 AI prompts
└── calc_perf.sh # 🚀 Performance calculation script
- Multi-Model Concurrency: Run multiple AI models simultaneously for trading
- Configuration Management: Support for JSON configuration files and environment variables
- Date Management: Flexible trading calendar and date range settings
- Error Handling: Comprehensive exception handling and retry mechanisms
| Tool | Function | API |
|---|---|---|
| Trading Tool | Buy/sell stocks, position management | buy(), sell() |
| Price Tool | Real-time and historical price queries | get_price_local() |
| Search Tool | Market information search | get_information() |
| Math Tool | Financial calculations and analysis | Basic mathematical operations |
- 📈 Price Data: Complete OHLCV data for NASDAQ 100 component stocks
- 📝 Trading Records: Detailed trading history for each AI model
- 📊 Performance Metrics: Sharpe ratio, maximum drawdown, annualized returns, etc.
- 🔄 Data Synchronization: Automated data acquisition and update mechanisms
- Python 3.10+
- API Keys: OpenAI, Alpha Vantage, Jina AI
# 1. Clone project
git clone https://github.com/HKUDS/AI-Trader.git
cd AI-Trader
# 2. Install dependencies
pip install -r requirements.txt
# 3. Configure environment variables
cp .env.example .env
# Edit .env file and fill in your API keysCreate .env file and configure the following variables:
# 🤖 AI Model API Configuration
OPENAI_API_BASE=https://your-openai-proxy.com/v1
OPENAI_API_KEY=your_openai_key
# 📊 Data Source Configuration
ALPHAADVANTAGE_API_KEY=your_alpha_vantage_key
JINA_API_KEY=your_jina_api_key
# ⚙️ System Configuration
RUNTIME_ENV_PATH=./runtime_env.json # Recommended to use absolute path
# 🌐 Service Port Configuration
MATH_HTTP_PORT=8000
SEARCH_HTTP_PORT=8001
TRADE_HTTP_PORT=8002
GETPRICE_HTTP_PORT=8003
# 🧠 AI Agent Configuration
AGENT_MAX_STEP=30 # Maximum reasoning steps# Install production dependencies
pip install -r requirements.txt
# Or manually install core dependencies
pip install langchain langchain-openai langchain-mcp-adapters fastmcp python-dotenv requests numpy pandas# 📈 Get NASDAQ 100 stock data
cd data
python get_daily_price.py
# 🔄 Merge data into unified format
python merge_jsonl.pycd ./agent_tools
python start_mcp_services.py# 🎯 Run main program - let AIs start trading!
python main.py
# 🎯 Or use custom configuration
python main.py configs/my_config.json{
"agent_type": "BaseAgent",
"date_range": {
"init_date": "2024-01-01", // Backtest start date
"end_date": "2024-03-31" // Backtest end date
},
"models": [
{
"name": "claude-3.7-sonnet",
"basemodel": "anthropic/claude-3.7-sonnet",
"signature": "claude-3.7-sonnet",
"enabled": true
}
]
}cd docs
python3 -m http.server 8000
# Visit http://localhost:8000| Rule Item | Setting | Description |
|---|---|---|
| 💰 Initial Capital | $10,000 | Starting capital for each AI model |
| 📈 Trading Targets | NASDAQ 100 | 100 top tech stocks |
| ⏰ Trading Hours | Weekdays | Monday to Friday |
| 💲 Price Benchmark | Opening Price | Trade using daily opening price |
| 📝 Recording Method | JSONL Format | Complete trading history records |
{
"agent_type": "BaseAgent",
"date_range": {
"init_date": "2025-01-01",
"end_date": "2025-01-31"
},
"models": [
{
"name": "claude-3.7-sonnet",
"basemodel": "anthropic/claude-3.7-sonnet",
"signature": "claude-3.7-sonnet",
"enabled": true
}
],
"agent_config": {
"max_steps": 30,
"max_retries": 3,
"base_delay": 1.0,
"initial_cash": 10000.0
},
"log_config": {
"log_path": "./data/agent_data"
}
}| Parameter | Description | Default Value |
|---|---|---|
agent_type |
AI agent type | "BaseAgent" |
max_steps |
Maximum reasoning steps | 30 |
max_retries |
Maximum retry attempts | 3 |
base_delay |
Operation delay (seconds) | 1.0 |
initial_cash |
Initial capital | $10,000 |
{
"date": "2025-01-20",
"id": 1,
"this_action": {
"action": "buy",
"symbol": "AAPL",
"amount": 10
},
"positions": {
"AAPL": 10,
"MSFT": 0,
"CASH": 9737.6
}
}{
"Meta Data": {
"2. Symbol": "AAPL",
"3. Last Refreshed": "2025-01-20"
},
"Time Series (Daily)": {
"2025-01-20": {
"1. buy price": "255.8850",
"2. high": "264.3750",
"3. low": "255.6300",
"4. sell price": "262.2400",
"5. volume": "90483029"
}
}
}data/agent_data/
├── claude-3.7-sonnet/
│ ├── position/
│ │ └── position.jsonl # 📝 Position records
│ └── log/
│ └── 2025-01-20/
│ └── log.jsonl # 📊 Trading logs
├── gpt-4o/
│ └── ...
└── qwen3-max/
└── ...
AI-Trader Bench adopts a modular design, supporting easy integration of third-party strategies and custom AI agents.
# Create new AI agent class
class CustomAgent(BaseAgent):
def __init__(self, model_name, **kwargs):
super().__init__(model_name, **kwargs)
# Add custom logic# Register in main.py
AGENT_REGISTRY = {
"BaseAgent": {
"module": "agent.base_agent.base_agent",
"class": "BaseAgent"
},
"CustomAgent": { # New addition
"module": "agent.custom.custom_agent",
"class": "CustomAgent"
},
}{
"agent_type": "CustomAgent",
"models": [
{
"name": "your-custom-model",
"basemodel": "your/model/path",
"signature": "custom-signature",
"enabled": true
}
]
}# Create new MCP tool
@mcp.tools()
class CustomTool:
def __init__(self):
self.name = "custom_tool"
def execute(self, params):
# Implement custom tool logic
return result- 🇨🇳 A-Share Support - Extend to Chinese stock market
- 📊 Post-Market Statistics - Automatic profit analysis
- 🔌 Strategy Marketplace - Add third-party strategy sharing platform
- 🎨 Cool Frontend Interface - Modern web dashboard
- ₿ Cryptocurrency - Support digital currency trading
- 📈 More Strategies - Technical analysis, quantitative strategies
- ⏰ Advanced Replay - Support minute-level time precision and real-time replay
- 🔍 Smart Filtering - More precise future information detection and filtering
- 💬 Discussions: GitHub Discussions
- 🐛 Issues: GitHub Issues
This project is licensed under the MIT License.
Thanks to the following open source projects and services:
- LangChain - AI application development framework
- MCP - Model Context Protocol
- Alpha Vantage - Financial data API
- Jina AI - Information search service
The materials provided by the AI-Trader project are for research purposes only and do not constitute any investment advice. Investors should seek independent professional advice before making any investment decisions. Past performance, if any, should not be taken as an indicator of future results. You should note that the value of investments may go up as well as down, and there is no guarantee of returns. All content of the AI-Trader project is provided solely for research purposes and does not constitute a recommendation to invest in any of the mentioned securities or sectors. Investing involves risks. Please seek professional advice if needed.
🌟 If this project helps you, please give us a Star!
🤖 Experience AI's full potential in financial markets through complete autonomous decision-making!
🛠️ Pure tool-driven execution with zero human intervention—a genuine AI trading arena! 🚀
Community Growth Trajectory
❤️ Thanks for visiting ✨ AI-Trader!