mathematics
Lean FRO and Mathlib Certify Alpöge’s Riemann Hypothesis Proof
Alpöge says he and Claude have a Riemann hypothesis proof. Most MIT Lean checks passed, but a comparator run is still in progress and there is no paper.
Latest
Anthropic Preprint Says a Claude Model Found a Way to Refute ETH
An Anthropic preprint says an internal Claude model found a 3-SAT algorithm that refutes the Exponential Time Hypothesis. It is unreviewed and asymptotic.
OpenAI Says Its AI Refuted the Exponential Time Hypothesis
OpenAI says its models refuted the Exponential Time Hypothesis and proved local smoothing in all dimensions. Neither claim has been peer reviewed yet.
Microsoft Says Human-Directed AI Agents Helped Prove KLS Conjecture
Microsoft Research says a preprint proves the 1995 KLS conjecture with human-directed AI agents. It is not peer reviewed; unguided runs fell short.
Simons, Anthropic and OpenAI Commit $15M for Maths AI Work
Simons Foundation partners with Anthropic and OpenAI on $15M to help mathematicians evaluate AI tools and set shared standards.
Preprint Claims Proof of Ghys’s Codimension-One Anosov Conjecture
A preprint credits Shanghai AI Lab's Intern-S2 with helping prove Ghys's codimension-one holomorphic Anosov conjecture; Lean formalisation is on GitHub.
AlphaProof Nexus Solved 9 of 353 Erdős Problems, DeepMind Says
DeepMind's AlphaProof Nexus is in Science. Its authors report 9 of 353 Erdős problems and 44 of 492 OEIS conjectures proved. What that means, and the limits.
OpenAI Publishes 372 AI Maths Results; About 42% Are Lean-Checked
Scientific American: OpenAI released 372 math/CS results from an unreleased model on 6 October 2026. Many are Lean-verified; independent check pending.
Timeline
Every story in mathematics, grouped by month.
November 2026
- 21 Nov 2026: Lean FRO and Mathlib Certify Alpöge’s Riemann Hypothesis ProofAlpöge says he and Claude have a Riemann hypothesis proof. Most MIT Lean checks passed, but a comparator run is still in progress and…
- 16 Nov 2026: GPT-6 Astra Helps Propose a Candidate Lattice Standard ModelIn an OpenAI science-blog guest post, Nathaniel Craig describes how GPT-6 Astra helped develop a candidate lattice definition of the Standard Model.
- 14 Nov 2026: Google Research Details Antigravity /teamwork Maths ResultsGoogle reports Antigravity /teamwork Lean formalisations, including Courtade–Kumar. Most results are preprints; Navier–Stokes is not claimed as solved.
- 10 Nov 2026: AlphaProof Code Formally Verifies Signal’s libsignal CodeDeepMind says AlphaProof Code verified compiled libsignal in a CLI harness with hooked primitives, generating over six million lines of Lean.
October 2026
- 28 Oct 2026: Axiom Says Its AI Agents Formalised the Simple Groups Proof in LeanAxiom says AI agents turned the 15,000-page Classification of Finite Simple Groups into a Lean proof in 24 days, and found 34 notable gaps…
- 28 Oct 2026: Anthropic Preprint Says a Claude Model Found a Way to Refute ETHAn Anthropic preprint says an internal Claude model found a 3-SAT algorithm that refutes the Exponential Time Hypothesis. It is unreviewed and asymptotic.
- 27 Oct 2026: OpenAI Says Its AI Refuted the Exponential Time HypothesisOpenAI says its models refuted the Exponential Time Hypothesis and proved local smoothing in all dimensions. Neither claim has been peer reviewed yet.
- 21 Oct 2026: Microsoft Says Human-Directed AI Agents Helped Prove KLS ConjectureMicrosoft Research says a preprint proves the 1995 KLS conjecture with human-directed AI agents. It is not peer reviewed; unguided runs fell short.
- 16 Oct 2026: Simons, Anthropic and OpenAI Commit $15M for Maths AI WorkSimons Foundation partners with Anthropic and OpenAI on $15M to help mathematicians evaluate AI tools and set shared standards.
- 15 Oct 2026: Preprint Claims Proof of Ghys's Codimension-One Anosov ConjectureA preprint credits Shanghai AI Lab's Intern-S2 with helping prove Ghys's codimension-one holomorphic Anosov conjecture; Lean formalisation is on GitHub.
- 10 Oct 2026: AlphaProof Nexus Solved 9 of 353 Erdős Problems, DeepMind SaysDeepMind's AlphaProof Nexus is in Science. Its authors report 9 of 353 Erdős problems and 44 of 492 OEIS conjectures proved. What that means,…
- 7 Oct 2026: OpenAI Publishes 372 AI Maths Results; About 42% Are Lean-CheckedScientific American: OpenAI released 372 math/CS results from an unreleased model on 6 October 2026. Many are Lean-verified; independent check pending.
Related stories
Latest from topics connected to mathematics. See all with connected topics
Latest on TSN
Explore the TSN Atlas: Every topic we cover, mapped. Click a bubble to dive in.
