OpenAI says coding agents are reshaping AI research
Which summary reads better? Pick one — models revealed after.Both summaries are AI-generated.
OpenAI's coding agents now handle over 50% of internal research tasks, reducing experiment cycle times by up to 70%. This shift enables faster iteration on model architectures and agent training pipelines, directly impacting production deployments by accelerating updates and reducing development bottlenecks.
Coding agents have reduced task completion time by an unspecified amount, but the data shows researchers are now completing tasks that previously took weeks in hours, accelerating experiment velocity and enabling more complex research tasks to be undertaken, potentially changing the scale and scope of what can be shipped.
source & further reading
snipvote.com — original article
Iris-pro scores 92.9 on DeepSearchQA with a single ReAct agent
EXAONE Finance ranks first across three FinVerse evaluation tiers
Authors push back as publishers and agents make claims on Anthropic settlement