InteractiveWorld WATCH

State-based checks can expose whether a multimodal agent actually changed the environment rather than narrated success. The non-commercial license, gated assets and high setup cost make it a watch item until cheaper reproduction or transfer evidence appears.

Next stepWait for lighter assets, an automated scoring protocol or real-device transfer results before deciding whether the reproduction cost is justified.

Judgment reviewed 2026-10-27 · Basis: Meta opens InteractiveWorld and a state-verifiable multimodal benchmark