IMO 2025 AI Solutions Comparison: OpenAI o3 vs Google DeepMind Gemini 2.5 Pro Deep Think

Live Project: imo-2025-ai-comparison.vercel.app

A comprehensive, side-by-side comparison of the official AI-generated solutions from leading research labs for the 2025 International Mathematical Olympiad (IMO).

🚀 Project Overview

The International Mathematical Olympiad (IMO) is the world's most prestigious mathematics competition for high school students. In 2025, AI models from OpenAI and Google DeepMind demonstrated unprecedented capabilities by solving complex IMO problems.

This project centralizes these solutions to facilitate analysis, discussion, and educational use.

Key Features:

• Side-by-Side Comparison: Direct view of OpenAI o3 and Google DeepMind Gemini 2.5 Pro Deep Think solutions.
• Full Coverage: Includes official solutions for Problems 1, 2, 3, 4, and 5.
• Clean Rendering: Custom LaTeX rendering using KaTeX for clear mathematical notation.
• Source Fidelity: Solutions are presented as close to the original model outputs as possible, ensuring authenticity.

⚠️ Disclaimer & Call for Contributions

This project is an open-source initiative managed by the community.

Please Note:

• Potential for Errors: While we strive for accuracy, the conversion of raw AI outputs (especially from OpenAI's non-standard formats) to web-friendly LaTeX can introduce transcription or formatting errors.
• Work in Progress: We are constantly refining the presentation and accuracy of the content.

We Need Your Help!
If you spot an error—whether it's a mathematical typo, a formatting glitch, or a missing detail—please contribute!

• Open an Issue: Let us know what's wrong.
• Submit a Pull Request: Directly fix the code or content.
• Join the Discussion: Help us improve the analysis of these AI achievements.

Your contributions ensure this resource remains accurate and valuable for everyone.

🛠️ Tech Stack & Setup

Built with simplicity and performance in mind:

• HTML5/CSS3: Clean, responsive design.
• Vanilla JavaScript: Lightweight logic for dynamic content loading.
• KaTeX: Fast, high-quality LaTeX mathematical rendering.
• Marked.js: Markdown parsing.

Local Development

To run this project locally:

1. Clone the repository.
2. Start a local server (e.g., Python):

   python3 -m http.server 8080

3. Open http://localhost:8080 in your browser.

📚 Sources

Solving 5 out of 6 problems officially, marking a historic milestone in AI mathematical

...

📚 Sources

• DeepMind IMO 2025 Official Blog: Gemini Deep Think
• DeepMind IMO 2024 Blog: AlphaProof
• OpenAI IMO 2025 Proofs (GitHub)
• OpenAI IMO 2025 Announcement (X)
• DeepMind IMO 2025 Solutions (PDF)

Acknowledgments

This project was built with the assistance of AI tools to accelerate development and content structuring, demonstrating the power of human-AI collaboration.