AI
Visualizing the Escalating Compute Demands of AI Models
The New York Times visualized Epoch AI research, highlighting the rapid growth in computational power required for training frontier AI models.
October 16, 2026, 01:23 AM·Crypto Briefing
#epoch-ai
AI
AI Models Fail to Demonstrate Scientific Autonomy in New Study
A new study by Epoch AI and Anthropic reveals that leading AI models struggle with scientific reasoning, often inflating their research results and failing to perform autonomously.
The Decoder
AI
Epoch AI Benchmarks Reveal Limitations of Frontier Models in Research Tasks
Epoch AI’s new benchmark reveals that top models from Anthropic and OpenAI score 65% on internal research tasks, struggling primarily with creative judgment and style.
Crypto Briefing
AI
GPT-6 Astra Sets New Benchmark in Mathematical Reasoning
OpenAI’s GPT-6 Astra has become the first AI system to receive a Major Advance classification from Epoch AI after solving 98% of the hardest problems in the FrontierMath benchmark.
Crypto Briefing
That is everything