ResearchBenchmarksStructured video reasoning on Keys and Doors Strict OODFollow22.7Exact MatchTBS-20-0.9085.22111.3517.479Jun 11, 2026Evaluation ResultsMethodMethodLinksExact MatchTBS-20Budget=20Budget=202026.0622.7TBS-10Budget=10Budget=102026.068.3BoN-20Budget=20Budget=202026.060.7BoN-10Budget=10Budget=102026.060.3BoN-1Budget=1Budget=12026.060BoN-5Budget=5Budget=52026.060