rotative
the decoder · August 20, 2026

59% success rate untrained: Generalist's GEN-1.5 robot model demos only

generalist ai unveiled gen-1.5, a robotics model that learns tasks from a 3-12 second demo with 59% untrained success. this physical prompting approach removes the need for expensive robot retraining.

read in 简体中文 · 繁體中文 · 日本語 · 한국어

generalist ai revealed gen-1.5, a generalist robotics model that executes tasks like opening jars after processing a single video demo into a context window. the model demonstrated a 59% zero-shot success rate across ten tasks, rising to 83% with minimal fine-tuning. unlike traditional robot learning, the model emerged with chain-of-thought and simulation transfer capabilities after eight months of interaction pretraining. this signals a shift toward generalist robot models that can adapt to new physical tasks instantly, though results currently rely on internal testing.

read at source ↗

← rotative