ResearchBenchmarksStructured Video Reasoning on Maze OverallFollow90.3Exact Match PercentageTBS-2049.42860.03970.6581.261Jun 11, 2026Evaluation ResultsMethodMethodLinksExact Match PercentageNormalized Time CostTBS-20Budget=20Budget=202026.0690.34.42TBS-10Budget=10Budget=102026.06843.53BoN-20Budget=20Budget=202026.06796.25BoN-10Budget=10Budget=102026.0675.24.01BoN-5Budget=5Budget=52026.0667.72.58BoN-1Budget=1Budget=12026.06511