Introducing ACT-2 Preview The first robotics model to unify broad...
Tony Zhao@tonyzzhao
73 views
Jul 20, 2026
~2 min read
Advertisement
1
Introducing ACT-2 Preview
The first robotics model to unify broad generalization with high reliability. A single fine-tuning example can teach Memo a new behavior that generalizes.
Zero shot, real unseen homes, 99% success rate.
The first robotics model to unify broad generalization with high reliability. A single fine-tuning example can teach Memo a new behavior that generalizes.
Zero shot, real unseen homes, 99% success rate.
2:25
2
We completed the most rigorous generalization test to date. Across 785 trials in 31 unseen environments, ACT-2 achieved 99.1% success in laundry folding.
ACT-2 also achieved human-level fold quality: receiving an average rating of 4.72/5, with 98.3% earning four or five stars.
ACT-2 also achieved human-level fold quality: receiving an average rating of 4.72/5, with 98.3% earning four or five stars.
3
Tons of emergent behavior along the way:
- Picking up clothes off the ground
- Handling baby wear to 8XL shirts
- Robustness against adversarial disturbances and lighting
- Picking up clothes off the ground
- Handling baby wear to 8XL shirts
- Robustness against adversarial disturbances and lighting
3:46
4
We call this a Solve.
Progress in robotics is difficult to measure because demos vary by setup. Demo ≠ Solved.
A Solve declares two boundaries: scope and adaptation cost. Without both, 99% has no context.
Progress in robotics is difficult to measure because demos vary by setup. Demo ≠ Solved.
A Solve declares two boundaries: scope and adaptation cost. Without both, 99% has no context.
5
We found a general recipe for Solves: scale pretraining, then hill-climb with minimal in-house data.
For the first time, one fine-tuning example can teach a new behavior that generalizes. Below: 4 folding strategies, learned from one example each and tested on held-out setups.
For the first time, one fine-tuning example can teach a new behavior that generalizes. Below: 4 folding strategies, learned from one example each and tested on held-out setups.
0:34
6
Quantitatively, we measure the generalization gap as the difference between in-domain and out-of-domain performance.
As we scale up pretraining, the gap falls sharply. This makes in-house performance a reliable predictor of performance in the wild.
As we scale up pretraining, the gap falls sharply. This makes in-house performance a reliable predictor of performance in the wild.
7
This property allows us to hill-climb performance in our office, and trust those gains to hold in unseen homes
Our fleet of Memos runs in parallel to rapidly advance reliability, quality, and speed.
Left: fleet-scale improvement in-house
Right: Memo working across unseen homes
Our fleet of Memos runs in parallel to rapidly advance reliability, quality, and speed.
Left: fleet-scale improvement in-house
Right: Memo working across unseen homes
8
Laundry is our first Solve of many. Our recipe is so general that scaling data and compute gives us predictable improvements.
Unlocking one Solve accelerates the next Solve.
The same ACT-2 model is learning to vacuum, organize toys, zip clothing, and turn pants inside out.
Unlocking one Solve accelerates the next Solve.
The same ACT-2 model is learning to vacuum, organize toys, zip clothing, and turn pants inside out.
1:50
9
This fall, ACT-2’s first Solve enters homes through our Beta Program, the final step towards fully autonomous home robot deployment.
Full technical report: sunday.ai/blog/act-2-pre…
Full technical report: sunday.ai/blog/act-2-pre…
