GEN-1.5
Generalist AI's robot model that learns a task from one short video clip loaded into its context window.
GEN-1.5, unveiled in August 2026, takes a demonstration of three to twelve seconds and puts it into the model’s context window, the working memory it can draw on while acting. Generalist AI calls this a physical prompt. The robot then attempts the task with no training step at all, which is the whole point: showing replaces retraining, and seconds replace days.
The reported numbers are 59 percent average success across ten tasks such as opening a jar or taking money from a wallet, rising to 83 percent after ten training steps on five minutes of data. The model can chain two demonstrations into longer sequences and accept clips recorded in simulation. Worth keeping in mind that the tasks shown are short and simple, and all results are the company’s own.