Google DeepMind's LLM-Lean agent solves nine open Erdős mathematical problems using formal proof search
The system also proved 44 conjectures from the OEIS.
The paper in general is worth reading, it's focused on areas where the lean ecosystem is more mature.
The Erdős problems are these:

"Our most capable agent autonomously resolved 9 of 353 open Erdős problems at the per-problem cost of a few hundred dollars, proved 44/492 OEIS conjectures, and is being deployed in combinatorics, optimization, graph theory, algebraic geometry, and quantum optics research."
You know we hit AGI when the AI solves problems that the vast majority cannot remotely grok. Furthermore, I doubt anybody without a long formal education can grok what the hell was just solved. So those advocating for the end of higher education and research have a Dunning-Kruger syndrome.
Another 9 open Erdos problems solved, this time by DeepMind team. Interesting loop of LLM - Lean agents working autonomously, and only after it's verified formally, going through human review.