**Recent breakthroughs in AI mathematical reasoning have reshaped expectations around the 2026 IMO.** In 2025, both OpenAI’s experimental reasoning model and Google DeepMind’s Gemini Deep Think officially reached gold-medal standard by scoring 35/42 on the six problems, solving five perfectly while missing only the hardest. This one-point improvement over 2024’s silver-level result (28/42) demonstrated that frontier large language models, enhanced with verification-and-refinement pipelines and deeper reasoning traces, could now match top human contestants under IMO conditions. Progress accelerated further into 2026 with additional systems from Chinese labs achieving perfect 42/42 scores shortly after the July competition in Shanghai. Traders pricing “No” at 69.5% appear to weigh remaining uncertainties around exact resolution criteria—such as whether solutions must be produced under identical time limits, without human curation, or in a fully public, non-experimental setting—against the swift capability gains. Key upcoming catalysts include any post-competition clarifications on grading standards and whether additional models meet the precise thresholds before market resolution.