This o3-mini Agent Can Predict Gold Prices!

Have you been tracking gold prices as of late? What if there was an agent that could autonomously send you daily gold rates? What if the agent could even predict tomorrow’s gold prices? With o3-mini AI integration, we can develop intelligent systems that can analyze financial trends and generate structured insights. In this article, we will build a price prediction agent with o3-mini to track 24-karat gold prices. It would leverage o3-mini function calling for accurate data retrieval and o3-mini’s structured output for clear, actionable predictions.

Why Should We Consider Using o3-mini?

There are so many different tools and platforms available today for building AI agents. Then why are we using o3-mini? Well, here are the main reasons for choosing o3-mini for building our price prediction agent:

• Efficient Automation & Real-Time Performance: Enables fast response times with reasonable accuracy, making it ideal for chatbots, coding assistants, and AI automation tools.
• Lightweight & Cost-Effective: Runs smoothly on consumer-grade hardware, reducing cloud costs and energy consumption.
• Strong Coding & Reasoning: Excels in debugging, code generation, and problem-solving despite its smaller size.
• Developer-Friendly & Scalable: Supports AI integration into workflows and deploys easily on mobile and edge devices.
• Versatile & Ideal for Experimentation: Enables rapid prototyping for AI-driven applications across automation, coding, and conversational AI.

To understand just how good this model is, here’s a comparison of o3-mini against other top models on various standard benchmarks.

Also Read: Is OpenAI’s o3-mini Better Than DeepSeek-R1?

And here’s how cost-effective it is:

Sources:

Setting Up the Environment

Before building the agent, we will first cover the necessary prerequisites and set up the environment.

Prerequisites

Accessing the API

We will be using two APIs here. One OpenAI API is used to access the o3-mini model, and the other Serper API is used to access the Google search engine to check today’s gold price.

Also Read: How to Run OpenAI’s o3-mini on Google Colab?

Building an AI Agent Using o3-mini

In this section, I will walk you through the process of building a gold price prediction agent using o3-mini. We will create an agent that analyzes market trends and provides insights into potential price movements.

I will guide you through making API calls, structuring prompts, and processing the generated predictions efficiently. By the end, you’ll have a working gold prediction agent capable of offering data-driven gold price forecasts. So let’s begin.

Now that you have both APIs, let’s move to the code editor and build our application.

Step 1:  Setting up Libraries and Tools to Guide AI Agent

First we will be importing all the necessary libraries and tools that we will need here

import os
import pandas as pd
from datetime import datetime, timedelta
from crewai_tools import ScrapeWebsiteTool, SerperDevTool
from crewai import Agent, Task, Crew, Process
from langchain_openai import ChatOpenAI
import warnings
warnings.filterwarnings('ignore')

search_tool = SerperDevTool()
scrape_tool = ScrapeWebsiteTool()

1. search_tool: The SerperDevTool() will help us to access the google search engine to retrieve the current information in our case its gold price.
2. scrape_tool: The ScrapeWebsiteTool() will help us to extract the relevant information from the searched webpage.

Step 2: Calling the API

To use the o3-mini and the google search engine first we have to load the OpenAI and Serper API. So this block will help us to loading the APIs.

with open("/home/vipin/Desktop/api_key/serper_api") as file:
    get_serper_api_key = file.read().strip()

with open("/home/vipin/Desktop/api_key/api_key_pro_blog.txt") as file:
    api_key = file.read().strip()
    
os.environ["OPENAI_API_KEY"] = api_key
os.environ["OPENAI_MODEL_NAME"] = "o3-mini-2025-01-31"
os.environ["SERPER_API_KEY"] = get_serper_api_key

Step 3: Creating Agents and Tasks

First, let’s create the required agents and define their tasks.

Creating Agents:

We will need 2 agents:

1. Gold Price Analyst agent will analyze the current price of gold in the market. Extracts structured data on gold prices and identifies macroeconomic factors influencing price movements.
2. Price Predictor agent will forecast whether the gold price will rise or fall based on historical and macroeconomic data and use past price trends, economic indicators, and machine learning insights to make daily predictions.

Gold Price Analyst

The Gold Price Data Analyst Agent gathers real-time gold prices and analyzes economic factors such as inflation, currency fluctuations, and global market trends. It provides structured insights that help refine market predictions.

gold_price_analyst_agent = Agent(
    role="Gold Price Data Analyst",
    goal="Continuously monitor and analyze 24K gold price in India from {url}.",
    backstory="This agent specializes in commodity price tracking, extracting structured data from {url}, and analyzing macroeconomic trends affecting gold prices.",
    verbose=True,
    allow_delegation=True,
    tools=[scrape_tool, search_tool]
)

Price Predictor

The Gold Price Predictor Agent utilizes machine learning models to forecast future gold price movements on a daily basis. It processes historical data and macroeconomic insights from the Analyst Agent to generate accurate predictions, continuously refining its model based on actual market trends.

gold_price_predictor_agent = Agent(
    role="Gold Price Predictor",
    goal="Predict whether 24K gold price in India will go up or down tomorrow using past data and macroeconomic trends.",
    backstory="Using historical data, macroeconomic trends, and machine learning insights, this agent predicts daily gold price movements and backtests predictions for accuracy.",
    verbose=True,
    allow_delegation=True,
    tools=[scrape_tool, search_tool]
)

Assigning the Tasks:

Now let’s define the tasks.

Task 1: Fetch Today’s Gold Price

The Fetch Today’s Gold Price Task is responsible for retrieving real-time gold prices in India from various global sources. It provides a detailed report that includes historical comparisons, regional price variations, and current market rates.

# Task for Data Analyst Agent: Analyze Market Data
fetch_todays_gold_price = Task(
    description="Scrape the latest 24K gold price in India from {url} and analyze macroeconomic trends affecting price movements.",
    expected_output="Today's gold price and key macroeconomic trends influencing its movement.",
    agent=gold_price_analyst_agent,
)

Task 2: Gold Price Predictor

The Gold Price Predictor Task leverages historical data, macroeconomic indicators, and machine learning models to forecast gold prices for different timeframes—next day, next week, next month, and next year. The output is a detailed report with future price predictions, enabling investors and financial analysts to plan their strategies with greater confidence.

gold_price_predictor = Task(
    description=(
        "Gather gold price data for the past 2 weeks and predict daily movements also give the sources from where you are taking the data. Mention the date and also the source of data taken"
        "Only look for last 3 days data while making prediction for any date. Validate each day's prediction against actual values, improving accuracy iteratively."
        "Don't look on actual value for each day before making prediction."
    ),
    expected_output="A detailed report on prediction accuracy over the past 2 weeks and tomorrow's expected gold price movement.",
    agent=gold_price_predictor_agent,
)

Step 4: Creating the Crew

The Gold Price Predictor Crew is a structured AI system that brings together all the agents. These agents work collaboratively to track real-time gold prices, analyze market influences, and predict future trends. Using the o3-mini model with a hierarchical process, the crew ensures an efficient workflow for gathering data, identifying fluctuations, and generating gold price forecasts.

gold_price_predictor_crew = Crew(
    agents=[gold_price_analyst_agent, gold_price_predictor_agent],
    tasks=[fetch_todays_gold_price, gold_price_predictor],
    manager_llm=ChatOpenAI(model="o3-mini-2025-01-31", temperature=0.3),
    process=Process.sequential,
    verbose=True
)

Step 5: Kickoff the Crew

The Gold Price Predictor Crew is initiated using the kickoff function with specific inputs. This triggers the collaborative workflow, where each agent executes its assigned task. The final result is a well-structured report, providing real-time prices, market insights, and future price predictions to aid investors in making informed decisions.

inputs = {"Query": "What will be the gold price movement in India for the next day?",
         "url": "https://www.goodreturns.in/gold-rates/"} 
result = gold_price_predictor_crew.kickoff(inputs=inputs)

Output

This will help us present the output in a markdown(.md) format.

from IPython.display import Markdown
markdown_content = result.raw
display(Markdown(markdown_content))

Caution!

As you can see in the output, the agent makes its predictions based on the majority of the outcomes of the previous days. Its similar to how we humans guess the next day’s price based on whether the price has been going up or down off late. This isn’t really accurate and shouldn’t be considered a robust price prediction tool. There are other machine learning models like ARIMA, LSTM, etc. that maybe more reliable in this case.

Hence, this analysis should only be considered as a demonstration to show how an agent for this task could work, and more importantly, to learn how to build an agent using o3-mini. Therefore, do not take it as financial advice. Financial markets are unpredictable, and trading them without proper due diligence is risky. Always conduct your own research and consult a financial professional before making investment decisions.

Conclusion

In this article, we’ve learned how to build a gold prediction agent using o3-mini, a fast and efficient AI model. o3-mini is a lightweight yet powerful tool for analyzing market trends, making it ideal for forecasting-related tasks. Its affordability and efficiency make it a great choice for developers looking to integrate AI into their workflows and strategies.

As gold and e-gold investments grow in popularity, having an AI-driven prediction agent can help identify profitable opportunities, reduce risks, and improve decision-making. By analyzing trends, the agent streamlines market research and provides timely insights. However, it’s important to note that this basic-level agent should not be solely relied upon for investment decisions. Always supplement AI insights with comprehensive market analysis and professional guidance.

Learn all about OpenAI’s o3-mini in this comprehensive course by Analytics Vidhya. Explore its core features and strengths and find out how it performs on various tasks through practical hands-on exercises.

Frequently Asked Questions

Hello! I’m Vipin, a passionate data science and machine learning enthusiast with a strong foundation in data analysis, machine learning algorithms, and programming. I have hands-on experience in building models, managing messy data, and solving real-world problems. My goal is to apply data-driven insights to create practical solutions that drive results. I’m eager to contribute my skills in a collaborative environment while continuing to learn and grow in the fields of Data Science, Machine Learning, and NLP.