Artificial intelligence systems can now remain autonomously focused on complex software tasks for nearly 15 hours at a time, according to new measurements from METR.
The research group estimates that Claude Opus 4.6 has a 50% time-horizon of about 14.5 hours on software tasks, meaning the model can complete half of the tasks that would take a skilled human that long without losing effectiveness.
Back in December, METR said Claude Opus 4.5 had a 50% time horizon of approximately 4 hours and 49 minutes on complex tasks.
The model evaluation group says the new figure is the highest point estimate it has reported so far, but notes that it comes with substantial uncertainty.
“We estimate that Claude Opus 4.6 has a 50%-time-horizon of around 14.5 hours (95% CI of 6 hrs to 98 hrs) on software tasks. While this is the highest point estimate we’ve reported, this measurement is extremely noisy because our current task suite is nearly saturated.”
The researchers note that the tasks are getting too easy for the newest models, which makes the measurement less precise. Because of the noise, the researchers say the confidence range is wide, from about 6 hours to 98 hours, and the upper end is actually longer than any task they tested.
“Near-saturation of the task suite can have unintuitive consequences for the time-horizon estimates. For example, the upper bound of the 95% CI is much longer than any of the tasks used for the measurement.”
METR’s chart shows steady growth in model time horizons over successive releases, with a best-fit trend implying a doubling roughly every 123 days based on recent data.
Box CEO Aaron Levie chimes in, predicting that the capability of AI for sustained work on software tasks will likely expand to other domains soon.
“The latest AI models are capable of long-running tasks that would have taken humans hours to do. The past year has seen a 15x increase in task length, which is insane. This is happening first in coding, but will cross the chasm into broad knowledge work.”
Disclaimer: Opinions expressed at CapitalAI Daily are not investment advice. Investors should do their own due diligence before making any decisions involving securities, cryptocurrencies, or digital assets. Your transfers and trades are at your own risk, and any losses you may incur are your responsibility. CapitalAI Daily does not recommend the buying or selling of any assets, nor is CapitalAI Daily an investment advisor. See our Editorial Standards and Terms of Use.