Research
InteractiveWorld benchmark tests whether MLLMs can guide humans through physical tasks
The benchmark has 120 multi-turn tasks across six categories, scored by state-based rewards. Across 12 frontier MLLMs, the strongest model, Claude Fable 5.1, reaches at most 20% on long-horizon Guidance tasks and 25% on Planning tasks.
Every outlet that ran it 1 article
Hugging Face PapersResearchLead