Meta
What Meta's Muse Spark maths papers mean for you
Meta says mathematicians used its Muse Spark 1.1 and 1.2 chat models to help write six peer-reviewed papers, five of which answer open questions.
The answer
Meta published six maths papers co-written with Muse Spark, five answering previously open research questions.
Meta has shown that its AI models can help professional mathematicians produce new results. On 2 October 2026, the company released six mathematics papers written with Muse Spark 1.1 and 1.2. Five of them answer research questions that were previously open.
What happened
According to Meta's research blog, the mathematicians worked with the models in Thinking Mode through the regular meta.ai chat. Meta said there was no custom research scaffold, meaning no specially built software around the model.
All six papers were peer reviewed, which means other experts checked them. Each paper marks which passages were drafted by the AI and which by the researchers.
The papers cover:
- Probability: "The Strict Threshold for Gaussian Ellipsoid Fitting" finds a sharp threshold for fitting random Gaussian points to an ellipsoid in high dimensions.
- Differential equations: finite-time blow-up (wave collapse) for radial negative-energy solutions of the mass-critical biharmonic nonlinear Schrödinger equation.
- Group theory: "Semiabelian Groups Need Not Be Monomial" gives a 384-element counterexample that disproves a 2024 conjecture by Kida.
- Optimisation: tightness of the cycle-based relaxation for completed length-three alpha-cycles.
- Arithmetic physics: a string two-point function equals a height function on a curve, linking number theory and p-adic string theory.
- Non-associative algebra: a three-dimensional counterexample disproves a conjecture by García-Martínez and Pérez-Rodríguez on solvable evolution algebras.
Meta said the models developed proof strategies, worked through calculations, wrote search programs and drafted technical sections. Humans set the direction and verified the results.
The blog describes the process this way: "Making progress means trying and retrying ideas, making new mistakes and resolving them, and sometimes going back to the beginning." The quote is from the Meta research blog.
What it means for you
The notes do not say anything about changes to Muse Spark's price, features or access, so there is nothing new to report on that front. What is clear is that the work was done in the same chat product the public uses, not in a separate lab system.
The papers also show a division of labour. The AI helped with strategy, calculation and drafting, and people checked the results. The marked passages in each paper let you see which parts came from which.
Meta has also said Muse Spark earlier reached gold-medal level on five high-school olympiad competitions. Muse recently reached #1 on the App Store, and OpenAI and Google DeepMind have also published AI-assisted maths results this year.
What happens next
Meta has not said in the material available what it plans next for this work. The six papers are the published record for now, and each one marks the AI's contribution so readers can judge it.
Sources
- Solving open research problems together — Meta AI, 2 October 2026