Epoch AI
Research organization studying trends in AI with data
Records 7
Newest first. Open one to see it with its source.
Large-scale errors were found even in a top-difficulty benchmark, raising doubts about the reliability of its review once again
A quantitative forecast that reasoning training will reach the frontier of total training compute within about a year, after which its growth rate will converge to around 4× per year
The first quantitative measurement breaking down, task by task, the steep drop in token prices at equal performance
Primary material on the problem of benchmark independence
A holdout design that cannot be contaminated
Estimated 400 trillion to 20 quadrillion effective tokens by 2030, and laid out the order of constraints: power binds first, then chip production capacity
The quantitative reference point for the data-exhaustion discourse