Infrathon tests language agents on long-horizon AI infrastructure tasks
Across 150 runs and over 10,000 GPU-hours, 135 still improved in the second half of a 24-hour budget.
1 stories, newest first · one card per story
Across 150 runs and over 10,000 GPU-hours, 135 still improved in the second half of a 24-hour budget.
Today 504 readers · 570 views · 181 reads · all time 187K views