Get the app
Robotics

Figure's humanoid tidied 30 homes it had never seen before

Figure's Helix 2.5 did chores in 30 Bay Area homes with no retraining. It succeeded 56% of the time, against 9% for a model trained from scratch.

Figure's new Helix 2.5 model got a humanoid robot doing chores in 30 Bay Area homes it had never entered. There was no fine-tuning, no data collection and no setup in any of them. It tidied living rooms, folded towels and made beds, and none of the toys, towels or bedding it handled appeared in its training data.

The headline number is more modest than the demo video: the robot succeeded 56% of the time. A trial only counted if the whole task was finished within strict time limits, and a human stepping in for safety counted as a failure. The same tasks trained from scratch managed just 9%. Figure also says Helix 2.5 matched Helix 02's success rate with half the task-specific data, and Helix 02 had been trained on data from the environment it was tested in.

The gain comes from pretraining on Index, Figure's dataset of recorded human behavior. It adds roughly 35 minutes of human experience every second. The robot learns how people move and handle objects before it is taught a specific chore. Figure itself says this shows promise, not that general-purpose home robots are solved.

Why it matters: a robot that works in a home it has never seen, without retraining, is what household humanoids need to be commercially viable. At 56%, it isn't reliable enough yet.

Sources

Written by an AI pipeline from the sources above. How it works.

The daily AI brief, on your phone.

Feed, daily deep-dive and bytes — readable offline, with push alerts for the topics you follow.

Get it on Google Play