Researchers developed a memory-efficient, single-GPU training and inference architecture for autoformalizing natural language mathematics into Lean 4 statement code, using the Qwen3-Coder-30B-A3B spar…
The Ethereum Foundation Formal Verification team, in collaboration with Yukon and zkSecurity, launched better.codes, an open autoresearch challenge that uses AI agents to raise the machine-checked sou…
A new pipeline called AViD Journal, described in an arXiv paper (arXiv:2608.14669v1), aims to automate novelty verification of mathematical theorems by formalizing LaTeX statements in Lean 4 and using…
MathCode, a terminal-based AI agent from math-ai-org, reduces Lean 4 compile-check latency from approximately 30 seconds to 0.4 seconds after warmup by using a persistent language server, enabling rea…
MathCode, an open-source terminal AI coding assistant from the Math-AI research community, converts plain-language math problems into Lean 4 theorems and attempts formal proofs, with a project page da…
MathCode, a terminal AI coding assistant with a built-in math formalization engine, automatically converts plain-language math problems into Lean 4 theorems and attempts formal proofs, featuring a per…
Researchers introduced Vero, the first benchmark to evaluate joint implementation and proof synthesis at the repository level, containing 43 multi-module instances from real-world repositories in Pyth…
A new arXiv paper (2608.10288v1) introduces Power Law Graph Attention (PLGA), a learned attention mechanism that exactly generalizes scaled dot-product attention (SDPA) and is used in the Power Law De…
A second AI system has produced a candidate improvement to the 67.25% zeta-zero bound, refining Anthropic's Theorem D. The new repository packages a stronger argument with an exact interval-arithmetic…
OpenAI's Astra AI system solved ten open mathematical problems, including the first explicit construction of a non-sofic group, a problem open since 1999, for a total compute cost of approximately $2,…
OpenAI released a 249-page manuscript claiming its internal Astra model solved ten open problems in mathematics and theoretical computer science, with Lean 4 formal proofs published on GitHub under Ap…
OpenAI released a 249-page math manuscript, ten machine-verified Lean 4 proofs, and a GitHub repository under Apache 2.0, introducing its next major model family, Astra, which solved ten decade-old ma…
A new arXiv paper (2607.28632v1) introduces a three-stage pipeline using large language models to systematically discover major mathematical conjectures, with experiments showing 20 out of 20 candidat…
OpenAI Group PBC announced Saturday that an internal version of its Astra model family solved 10 long-open problems in mathematics and theoretical computer science, publishing machine-checkable Lean 4…
A developer has proposed an architectural pattern that combines Lean 4 formal verification with ClickHouse real-time analytics to create verifiable AI infrastructure. The approach uses Lean 4's depend…
OpenAI researcher Noam Brown announced on August 1 that an internal version of Astra, OpenAI's next major model family, solved ten open problems in mathematics and theoretical computer science, includ…
OpenAI on Saturday published ten claimed advances in mathematics and theoretical computer science, generated by an internal version of its forthcoming Astra model, including an explicit non-sofic grou…
On July 25, Ramana Kumar published a repository containing an AI-assisted 'disproof' of the Collatz conjecture that compiled in Lean 4 and was accepted by the independent checker nanoda, but on July 2…
OpenAI published ten results on long-standing mathematical problems, including high-dimensional sphere packing and Connes's rigidity conjecture, found by an internal model called Astra at a token cost…
Researchers introduced CircuitProver, an agentic Lean 4-based verification framework that automatically translates parameterized hardware designs and natural language specifications into executable Le…