OpenAI and Google’s AI Models Score Gold at 2025 Math Olympiad — But the Rivalry Heats Up

AI Systems Match the Best Young Math Minds — and Each Other

Artificial Intelligence just hit another major milestone. Both OpenAI and Google DeepMind revealed that their AI models achieved gold medal-level scores in the 2025 International Math Olympiad (IMO) — one of the world’s toughest math competitions for high school students.

This achievement highlights how fast AI systems are evolving. But it also shows something more subtle: the fierce, close competition between leading AI labs.

Key Takeaways

• OpenAI and Google’s AI models both solved 5 out of 6 IMO problems — a gold-medal performance.

• Both used informal systems that read and answered math problems in natural language.

• Google waited for official IMO approval before sharing results.

• OpenAI went public sooner, using third-party evaluators.

• Debate aside, the real story is how fast AI is improving at reasoning and problem-solving.

What’s the Big Deal About IMO?

The International Math Olympiad is more than just a student contest. It’s a benchmark for high-level reasoning — a skill AI systems have traditionally struggled with. Winning here isn’t about simple calculations; it’s about proving complex ideas logically.

In 2024, Google’s AI earned a silver medal, but it needed humans to reformat questions. This year, things changed.

Both OpenAI and Google used systems that could:

• Read natural-language math problems, like a student would.

• Generate step-by-step proof-based answers on their own.

• Achieve results on par with top human contestants.

Gold Medals, but No Clear Winner

OpenAI and Google both claim success. But they took different routes to get there — and that’s where the controversy begins.

What OpenAI Did

• Used a natural-language model trained for reasoning.

• Brought in three former IMO medalists to grade the answers.

• Learned it had gold-level performance.

• Reached out to IMO for guidance — and was told to wait.

• Announced its results anyway, shortly after the student award ceremony.

What Google Did

• Partnered with IMO organizers months in advance.

• Waited for official grading and approval from IMO’s president.

• Shared results after the student awards, following protocol.

A Battle of AI… and Optics

Let’s be honest — both models performed incredibly well. But this isn’t just about math. It’s also about who looks more trustworthy, more advanced, and more respectful of the human competition.

That matters because:

• The AI race is about more than product features — it’s about public perception.

• Top AI researchers often come from math backgrounds, and many follow events like IMO closely.

• These “vibes” influence recruitment, funding, and partnerships.

And right now, the vibe is that Google and OpenAI are neck-and-neck in high-stakes AI development.

What the Results Really Show

OpenAI and Google both reached gold-medal scores. That’s rare — only a small percentage of students reach that level at IMO. These AI systems didn’t just pass the test. They excelled.

Even more impressively, they did it without handholding. No humans needed to translate problems into a format AI could understand. The models read, understood, and solved the questions directly.

This shows big progress in:

• AI reasoning in ambiguous and proof-heavy domains.

• Handling non-verifiable tasks, like research or high-level decision-making.

• Natural language understanding, without strict formatting or cues.

What’s Next in the AI Race?

OpenAI once seemed far ahead. Now, it’s clear that Google is catching up — fast.

With GPT-5 expected soon, OpenAI may regain its edge. But Google is showing it can match performance and follow formal channels when it matters.

Either way, this rivalry is pushing both companies — and AI technology — forward.

Summary: Why It Matters

Point	Summary
AI Progress	Both models showed elite-level problem solving in natural language.
New Standard	No translation needed — just AI reading and solving complex math like a student.
Competitive Edge	OpenAI went fast with informal grading; Google took the formal, official route.
Perception Battle	Public trust, researcher respect, and transparency are as important as performance.
Future Outlook	Expect more breakthroughs — and more drama — as GPT-5 and DeepMind’s next moves unfold.

Final Thoughts

This isn’t just a story about math. It’s about how fast AI is learning to think, how closely matched the leaders are, and how even a math contest can turn into a PR showdown.

While OpenAI and Google debate gold medals, the rest of the world should take note: AI just aced one of the hardest tests we have — and it’s only getting smarter.

this post was originally published on https://techthrilled.com/press-release/openai-google-math-olympiad-2025/