Epoch AI

Epoch AI

Home
Gradient Updates
The Epoch Brief
Epoch After Hours
Archive
About

Beyond benchmark scores: Analyzing o3-mini’s mathematical reasoning

READ THE LATEST

Recent posts

View all
GPQA Diamond: What’s Left?
Investigate GPQA Diamond benchmark’s validity: uncover flawed questions, model challenges, and why it still informs AI evaluation.
 • 
Greg Burnham
Is AI already superhuman on FrontierMath?
How do humans and AIs compare on FrontierMath? We ran a competition at MIT to put this to the test.
 • 
Anson Ho
How Fast Can Algorithms Advance Capabilities?
This week’s issue is a guest post by Henry Josephson, who is a research manager at UChicago’s XLab and an AI governance intern at Google DeepMind.
How far can reasoning models scale?
Available evidence suggests that rapid growth in reasoning training can continue for a year or so.
 • 
Josh You
Where’s my ten minute AGI?
Why don’t AIs automate more real-world tasks if they can handle 1-hour ones? Anson Ho explores key capability and context bottlenecks.
 • 
Anson Ho
The case for multi-decade AI timelines
In this Gradient Updates weekly issue, Ege discusses the case for multi-decade AI timelines.
 • 
Ege Erdil
Epoch AI
Epoch AI
Weekly commentary on AI news and developments.
Read more from Epoch AI
Visit our website
Newsletters
Logo for Gradient Updates
Gradient Updates
Gradient Updates is a newsletter by Epoch AI, providing shorter-form commentary on important issues around AI.
Logo for The Epoch Brief
The Epoch Brief
Our most recent research and publications.
Logo for Epoch After Hours
Epoch After Hours
A podcast where Epoch AI researchers discuss their thoughts and candid takes on the present and future of AI.
Podcasts
Coming soon
Get the latest from Epoch AI
© 2025 Epoch Artificial Intelligence, Inc
Substack
Privacy ∙ Terms ∙ Collection notice
Start writingGet the app
Substack is the home for great culture