Meta claims its Muse Spark model helped mathematicians solve five open problems. The results look real, but the story is more nuanced than the headline.
On October 2, 2026, Meta AI published six math papers written together with human mathematicians using Muse Spark 1.1 and 1.2 in “Thinking Mode.” According to the company, five of the six answer questions that were open. The papers cover probability, partial differential equations, group theory, optimization, arithmetic physics, and non-associative algebras.
Meta says the mathematicians worked through the regular meta.ai chat interface, with no custom research scaffolding. Each paper states which sections were drafted by the AI and which by humans, and all went through review by other mathematicians.
The human authors named in the coverage include Aykut Arslan, Leonard Dinh, Joseph Phillip Brennan, and Milana Golich. The role described is that of an assistant that generates search code, proposes proof strategies, and drafts sections under expert direction, not that of an autonomous discoverer.
That's why the “five open problems” figure should be read as Meta's claim, not a settled fact. What is well documented is that a general-purpose model, from a chat window, produced a verifiable counterexample and useful proof proposals.
Carlos Montiel is an enterprise AI solutions architect. He implements LLMs, Agents, RAG, and orchestrators for companies across Guatemala and Latin America. Reach out for a consultation.
Contact Carlos Montiel