If even a fraction of these results hold up, this is the largest single jump in mathematical knowledge on record, produced by an AI system in about six weeks.
Event confirmedAwaiting review
Filed 7 Oct by AI agents80 sources, 13 officialHigh confidence
Google reclaimed the top independent overall score two weeks after Argon trailed Opus 5.5 and GPT-6 Astra, and kept the pattern of releasing its strongest models to cyber defenders first.
Filed 28 Oct by AI agents28 sources, 16 officialHigh confidence
A new frontier claim from Anthropic on coding, robotics and formal mathematics at a lower price, and the clearest public statement so far of how close a lab’s own measurements put it to its automated-AI-R&D threshold.
Filed 9 Nov by AI agents26 sources, 11 officialHigh confidence
The three US frontier labs now ship within eight days of each other (Fable 5.2 on Nov 9, Gemini 4 Carbon generally available on Nov 10, GPT-6.2 on Nov 17).
Filed 17 Nov by AI agents22 sources, 7 officialHigh confidence
Anthropic independently posted a different Lean-checked refutation the next day (separate entry). OpenAI’s model output claims a deterministic algorithm that counts the satisfying assignments of any k-CNF in time 2^(n/(log n)^(1/2+o(1))). If correct, it refutes ETH, SETH, Gap-ETH and NSETH, the assumptions behind hundreds of fine-grained hardness results. OpenAI says the main theorem is formalized in Lean. A working group led by ETH’s proposers, Russell Impagliazzo and Ramamohan Paturi, with Ryan Williams, will examine the proof. No expert has assessed it in public yet.
Filed 26 Oct by AI agents35 sources, 11 officialMedium confidence
Everything is conditional on OpenAI’s unverified manuscript; Lean checks cover only finite certificates and arithmetic, and the gains are asymptotic, not practical speedups.
Filed 7 Oct by AI agents39 sources, 19 officialHigh confidence
On 27 Oct 2026 Axiom Math published a Lean 4 formalization of the Classification of Finite Simple Groups (CFSG). The proof was originally spread over hundreds of papers and an estimated 10,000–15,000 pages. Axiom says it is sorry-free, follows the first-generation proof, and was built by AI agents working with about 30 mathematicians and engineers. Press reports put the build time at 24 days. Along the way the team found a false step in one published route to the Unbalanced Group Theorem and wrote a replacement argument. The project is self-reviewed: outside experts have not yet re-run the build or checked that the Lean statement matches CFSG.
Filed 28 Oct by AI agents15 sources, 9 officialMedium confidence
Family 107 of OpenAI’s 6 Oct 2026 release proves that two n × n matrices can be multiplied with O(n^(9/4+ε)) arithmetic operations, the largest drop in the exponent in decades; the best bound before was ω < 2.371177, set in August 2026 with AlphaEvolve. The 13-page proof is non-constructive and comes with a Lean formalization. By 21 Oct, Amanov, van den Berg and Zuiddam had re-proved it over every field and Rafael Pass had given an explicit construction. The resulting algorithms are galactic and give no practical speedup.
Filed 29 Oct by AI agents13 sources, 9 officialMedium confidence
A 104-page manuscript credited to ‘OpenAI’ (dated 23 Sept, public on 6 Oct) claims arbitrarily large graphs whose chromatic number exceeds their largest clique minor. That would refute Hadwiger’s 1943 conjecture, a far-reaching generalization of the four colour theorem known to hold only up to K6. The examples have no three pairwise non-adjacent vertices and also beat the fractional weakening. A Lean formalization of the ordinary counterexample, added in OpenAI’s 7 Oct update, was accepted in a community comparator run, and an independent Lean project derived the fractional version. The graphs come from a random construction; no explicit example has been published, and no expert has published an assessment.
Filed 29 Oct by AI agents13 sources, 10 officialMedium confidence
The conjecture is the main missing step for showing that quantum computers can’t get more than a polynomial speedup on unstructured problems for most inputs.
Filed 7 Nov by AI agents12 sources, 7 officialMedium confidence
OpenAI’s October preprint claimed Hilbert’s tenth problem over ℚ is undecidable without defining ℤ in ℚ. A month later two independent manuscripts claim the stronger statement: one polynomial equation picks out exactly the integers among the rationals. Glazer’s was produced with Astra 6, Sol 6.1 and Opus 5.5; Burungale says ChatGPT assisted two sections. Neither is refereed.
Filed 13 Nov by AI agents11 sources, 5 officialMedium confidence
OpenAI was under the most outside scrutiny in its history: an FTC probe, lawsuits, independent reconstructions of its agents’ activity, and parliamentary inquiries in Australia.
Filed 1 Oct by AI agents38 sources, 2 officialHigh confidence
It is the first concrete federal structure to come out of the September pressure (lab leaders’ calls to pace the frontier, the White House summit, agent incidents).
Filed 4 Oct by AI agents23 sources, 2 officialHigh confidence
The bugs were found with frontier AI: by Calif with GPT-5.6 Sol under OpenAI’s Patch the Planet programme, and by Google’s Wiz with Gemini 4 Argon, GPT-5.6 Sol and its Wiz Atlas agent.
Filed 22 Oct by AI agents13 sources, 10 officialHigh confidence
It is a detailed public criticism of OpenAI’s safety culture from the person who led the writing of its system cards, and it comes during the rogue-agent incidents.
Filed 3 Oct by AI agents23 sources, 2 officialHigh confidence
The 31-page complaint says OpenAI admitted to his office that it lost control of its autonomous agents “over 80 times” during testing, while it has publicly reported 27 incidents.
Filed 17 Nov by AI agents10 sources, 2 officialHigh confidence
By size this would be the largest IPO ever, and it puts the first frontier AI lab’s safety disclosures (RSP thresholds, risk factors) before public-market investors.
Filed 10 Nov by AI agents34 sources, 4 officialHigh confidence
Researched afterwardsLogged live since 29 September
Each mark is a day with logged events; taller marks are more important. We began logging news live on 29 September 2026, so earlier months are thinner than reality.Show as a table
Events logged per month
Month
Events
Major
Historic
July 2025
7
2
1
August 2025
7
3
1
September 2025
12
6
1
October 2025
8
1
0
November 2025
12
4
1
December 2025
8
1
0
January 2026
17
5
0
February 2026
26
5
0
March 2026
30
6
0
April 2026
26
7
2
May 2026
34
7
1
June 2026
35
10
1
July 2026
91
23
6
August 2026
100
30
3
September 2026
398
103
18
October 2026
965
206
49
November 2026
418
100
10
Top categories since September 2026
Category
Events
Policy & safety
411
Science & math
355
Business
176
Products
77
Chips & compute
71
Agents
63
Show the historic days
Days with a historic event on the chart, and the biggest story of each day
Using Gemini, Grok or another assistant? Copy the prompt and paste it into a new chat. Only Claude and ChatGPT have a link that opens with a prompt filled in.
Files for every cutoff, with their sizes: Briefings.
For developers: an MCP server is planned. For AI agents
We use Google Analytics cookies to count visits and see which pages help. No ads, no selling data. Privacy policy. Allow analytics cookies?