{"type":"post","id":"interactiveworld-a-verifiable-human-centric-environment-for-building-ac7901","slug":"interactiveworld-a-verifiable-human-centric-environment-for-building-ac7901","title":"InteractiveWorld: A Verifiable Human-Centric Environment for Building Better Interactive MLLM Assistants","description":"A research team introduced InteractiveWorld, a simulated environment for testing multimodal assistants that guide people through physical tasks, and InteractiveBench, a 120-task benchmark built on it.","retrieval_nugget":"A research team introduced InteractiveWorld, a simulated environment for testing multimodal assistants that guide people through physical tasks, and InteractiveBench, a 120-task benchmark built on it. The paper reports that even its strongest tested model scored no higher than 20% on long-horizon guidance and 25% on planning, underscoring the gap between fluent responses and reliable help over time.","published_at":"2026-11-07","updated_at":"2026-11-07","record_date":"2026-11-07","date_kind":"published_at","topics":["claude"],"entities":["InteractiveWorld","InteractiveBench","Claude Fable 5.1","arXiv"],"related_patterns":["ai-engineering-evaluations-are-splitting-into-task-specific"],"novelty_label":"new_fact","what_changed":"InteractiveWorld introduces a human-centric, verifiable benchmark for interactive MLLM assistants, reporting that the strongest tested model scored at most 20% on long-horizon guidance and 25% on planning.","prior_refs":["https://newruntime.com/posts/aa-3d-ranks-21-models-by-human-votes-ae34ef/"],"source_url":"https://arxiv.org/abs/2610.33966","source_urls":["https://arxiv.org/abs/2610.33966"],"schema_version":"newruntime-agent-readable-v0.2","stable_id":"post:interactiveworld-a-verifiable-human-centric-environment-for-building-ac7901","status":"published","visuals":[],"editor":{"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]},"routes":{"html":"https://newruntime.com/posts/interactiveworld-a-verifiable-human-centric-environment-for-building-ac7901/","markdown":"https://newruntime.com/posts/interactiveworld-a-verifiable-human-centric-environment-for-building-ac7901.md","json":"https://newruntime.com/posts/interactiveworld-a-verifiable-human-centric-environment-for-building-ac7901.json"},"markdown_url":"https://newruntime.com/posts/interactiveworld-a-verifiable-human-centric-environment-for-building-ac7901.md","next_reads":[{"type":"topic","path":"/topics/claude/","url":"https://newruntime.com/topics/claude/","title":"Claude","reason":"Explore the claude topic hub.","media_type":"text/html","markdown_url":"https://newruntime.com/topics/claude.md","json_url":"https://newruntime.com/topics/claude.json"},{"type":"collection","path":"/posts/","url":"https://newruntime.com/posts/","title":"Field Notes","reason":"Open the latest editorial analysis.","media_type":"text/html","json_url":"https://newruntime.com/posts.json"},{"type":"collection","path":"/sources/","url":"https://newruntime.com/sources/","title":"Sources","reason":"See the original sources this archive cites.","media_type":"text/html"}]}