@EpochAIResearch: OpenAI recently published data on its researchers’ coding-agent usage. We took their weekly figures from January to mid…
Summary
Epoch AI analyzed OpenAI's published data on its researchers' coding-agent usage, finding that agent spending doubled roughly once a month between January and mid-August 2026.
View Cached Full Text
Cached at: 10/06/26, 09:47 AM
OpenAI recently published data on its researchers’ coding-agent usage. We took their weekly figures from January to mid-August 2026 and fitted trends to summarize how quickly spending grew.
We found that spending has been doubling about once a month. https://t.co/TpqyvhWqv9
Similar Articles
@simonw: Interesting details in here about how OpenAI researchers are using coding agents I'd love to know what caused the huge …
OpenAI's Kevin Liu shared data on how the company's researchers use AI coding agents to accelerate internal research, prompting discussion about recursive self-improvement and a sharp mid-July spike in token usage.
@OpenAI: Coding agents are helping scientists spend more time advancing research, taking on everything from routine maintenance …
OpenAI highlights how coding agents can assist scientists with research tasks such as routine maintenance and optimization, while emphasizing that researchers must still define scientific questions and verify results.
OpenAI: AI agents now perform 3.1 researcher-workdays for every human researcher-workday, says it has reached “automated research intern” level, and expects “automated AI researcher” by March 2028
OpenAI claims its AI agents now contribute 3.1 workdays for every human researcher workday, achieving an "automated research intern" level, and predicts it will reach an "automated AI researcher" by March 2028. The company says agentic systems have accelerated progress toward recursive self-improvement, effectively scaling its research workforce from roughly 1,000 to an equivalent of over 4,000.
Research acceleration: The view inside OpenAI
OpenAI reports reaching its goal of an automated research intern system and is making strong progress toward a fully automated AI researcher by March 2028, while detailing how internal coding agents are accelerating research and emphasizing the need for parallel alignment and safety work.
The insanity of 10.000 agents running
The article highlights the massive scale of OpenAI's effort in running 10,000 agents for 88 hours, nearly equivalent to a century of work, underscoring the potential impact of AI combined with abundant compute resources.