Figure ai ran its helix 2.5 humanoid model through 30 homes it had never seen and the robots tidied…
figure ai ran its helix 2.5 humanoid model through 30 homes it had never seen and the robots tidied living rooms, folded towels, and made beds without collecting new data or retraining. zero-shot generalization stopped being a lab claim on september 18.
the test that matters for home robots is not a benchmark, it is a stranger house.
Context
Figure's post of 17 September 2026 says it evaluated Helix 2.5 on three behaviors, tidying a living room, folding towels and making a bed, in 30 Bay Area homes, with no data collected in the evaluation homes. It says success required the whole task with no partial credit, and that a policy pretrained on its index succeeded on 56% of zero-shot trials against 9% for a from-scratch policy trained on the same task data. Figure writes that the point is not that general humanoid robotics is solved.
The results are Figure's own and not independent. No new data, fine-tuning or adaptation applied in the evaluation homes, but each behavior was specified by fine-tuning data collected elsewhere, so without retraining holds for the homes and not for the task. The note leaves out the 56% success rate, and the evidence is three tasks in Bay Area homes. The first demonstration is Figure's own claim, to its knowledge. The post is dated 17 September, and the note's 18 September was not found in the text read. Zero-shot stopped being a lab claim is the author's opinion.
Watch next
- Independent replication or third-party evaluation.
Sources
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 19 September 2026 at 02:33 IST. Sources are the papers and datasets the note draws on.
View the original post ↗Embed this note
More notes
The air is now being asked to keep its own ledger
the air is now being asked to keep its own ledger: ecmwf’s aifs compo becomes the first ai model to forecast atmospheric composition globally every three hours, cleanair simulates 365 days of pm2.5 over china in ten seconds, and a unified framework maps six pollutants at one kilometer across the whole country. the air now files its own composition report.
read the note →The current is now being asked to draw its own map
the current is now being asked to draw its own map: china’s langya 2.0 predicts six ocean phenomena including internal waves and mesoscale eddies, a deep net called wenhai resolves eddies globally with air sea flux formulas built in, and scripps infers surface currents from the way temperature patterns deform in satellite images. the ocean now files its own circulation report.
read the note →The soil is now being asked to report its own carbon
the soil is now being asked to report its own carbon: a nix color sensor paired with generative data augmentation predicts soil organic carbon without a lab, random forest drives 74 percent of soil health mapping studies, and sentinel 2 tracks five year carbon change across france and italy from 922 samples. the dirt now files its own carbon account.
read the note →