Epoch AI
Subscribe
Sign in
Home
Gradient Updates
Data Insights
Papers & Reports
Epoch After Hours
The Epoch Brief
Updates
Archive
About
Latest
Top
Discussions
GPT-5.1 is about as capable as GPT-5
With “high” reasoning, both GPT-5.1 and GPT-5 score 151 on the Epoch Capabilities Index, our tool for combining results across multiple benchmarks
11 hrs ago
•
Luke Emberson
The software intelligence explosion debate needs experiments
The existing debate rests on data and assumptions that are shakier than most people realize. To make progress, we need better evidence, and experiments…
Nov 15
•
Anson Ho
and
Parker Whitfill
17
2
3
Introducing the Frontier Data Centers Hub
We announce our new Frontier Data Centers Hub, a database tracking large AI data centers using satellite and permit data to show compute, power use, and…
Nov 4
•
Epoch AI
30
2
3
OSWorld — AI computer use capabilities
Tasks are simple, many don't require GUIs, and success often hinges on interpreting ambiguous instructions. The benchmark is also not stable over time.
Nov 4
•
Greg Burnham
15
1
October 2025
The Epoch AI Brief - October 2025
Report on decentralized training, new Epoch Capabilities Index for tracking AI progress, FrontierMath evaluations of leading models, revenue insights on…
Oct 31
•
Epoch AI
10
AI + Math Chat #2 - Formalizing Proofs & Breaking Codes
“Can LLMs go beyond pattern-matching to produce fully formalized mathematical proofs and novel cryptographic attacks?” w. Kevin Buzzard, Alex…
Oct 30
•
Epoch AI
5
2
1
AI + Math Chat #1 - Thinking smarter not harder
“How close are today’s large-language models to genuine mathematical creativity?” with mathematicians Richard Borcherds, Sergei Gukov, Daniel Litt, and…
Oct 30
•
Epoch AI
2
1
Solving AI’s power problem with decentralized training
Conventional wisdom in AI is that large-scale pretraining needs to happen in massive contiguous datacenter campuses. But is this true?
Oct 28
•
Jaime Sevilla
and
Anton Troynikov
17
2
AI can learn logic. But can it learn folklore knowledge? - Svetlana Jitomirskaya
Large language models can imitate reasoning steps and even verify formal proofs. But Svetlana Jitomirskaya argues they lack folklore knowledge: the…
Oct 27
•
Epoch AI
10
2
1
Waiting for AI's phase change in mathematics - Ravi Vakil
Stanford mathematician Ravi Vakil, president of the American Mathematical Society, expects AI’s impact on mathematics to come as a phase change, not a…
Oct 23
•
Epoch AI
9
1
Less than 70% of FrontierMath is within reach for today’s models
57% of problems have been solved at least once
Oct 17
•
Greg Burnham
11
2
OpenAI is projecting unprecedented revenue growth
No company has gone from $10B to $100B as fast as OpenAI projects to do
Oct 15
•
Greg Burnham
29
6
4
This site requires JavaScript to run correctly. Please
turn on JavaScript
or unblock scripts
