Human expert baseline attempt on an 8-hour AI R&D task (METR RE-Bench)
approximately $1,855 per expert attempt (8-hour task)
Dated 27 May 2025Source checked 11 Oct 2026Page out of reach at the last check, 17 Oct 2026Basis: stated
In the source's own words
“This is only a small fraction of the approximately $1,855 that we paid our human experts on average”
Where it was read
arXiv, RE-Bench (v2), checked 11 Oct 2026
Everything read from this source (1)Archived copy, 4 Oct 2026
What it is about
Figures it counts in
- $300 per taskDatasets, tasks and environments, paid to the people who make it, per task
- 9.5xHow much more buyers pay than makers are paid for datasets, tasks and environments, per task
- 36Prices per task behind every valuation of a task set