OpenAI Says It Hit 'Automated Research Intern' Milestone

OpenAI says it has met the 'automated research intern' milestone Sam Altman set in an October 2025 livestream, and disclosed in a Sept. 6 post, Research acceleration: The view inside OpenAI, that its research organization is now running 3.1 agent-workdays of effort for every workday of human labor, measured in mid-August.

'According to our measurements, we have now reached the goal, announced last fall, of having an automated research intern by September of this year,' the company writes, defining that intern as 'a system that can carry out well-defined research tasks under human direction, including tasks that would take a skilled researcher a few days.' Before June 2026, total agent runtime across the research org was still below total human labor.

The compute bill behind that shift is unusually specific. According to Engadget's read of the post, the median OpenAI researcher was running more than $600 a day of inference on internal agents at API prices by mid-August, with the heaviest users burning past $7,000 a day. Per Inside AI News, over half of the successful 4-8 hour agent tasks still required at least one human intervention.

OpenAI paired the milestone with a hedge on the direction of travel. 'We do not yet know how to safely get all the way to aligned, full RSI,' the post says. 'We cannot assume that progress in alignment and safety will keep pace.' The company adds that 'if it is done responsibly, we believe automated AI research will yield models that directly enhance human welfare and advance OpenAI's mission,' and says it is now targeting a fully automated AI researcher by March 2028, the second half of the schedule Altman floated in his October 2025 livestream. The post lands two days after the company's own sandbagging admission about GPT-6 Astra.