Meet Kyutai REKI: The 2.4B VLA Model Bringing Efficient Memory to Robot Manipulation

Kyutai, Inria, and École des Ponts released REKI, a 2.4 billion parameter vision-language-action model for robot manipulation. Trained only on public data, it uses a lightweight memory of recent observations and performs strongly on out-of-distribution benchmarks.

AI Pillar Hub files a short pointer. The reporting lives at the source, MarkTechPost.