List your company

ResearchFormerly WizWand

Path-level reasoning

Benchmarks

DatasetSOTA methodMetricTrendResultsLast updated

UNOBench synthetic Hard (test)

56.8SR-P

10

Feb 26, 2026

UNOBench synthetic Medium (test)

74.8SR-P (%)

10

Feb 26, 2026

UNOBench synthetic Easy (test)

82.8SR (Precision)

10

Feb 26, 2026

UNOBench synthetic No obstructions (test)

94.8SR (%)

10

Feb 26, 2026

UNOBench real (Hard)

79.5SR-P

10

Feb 26, 2026

UNOBench real (Medium)

76.6SR-P (%)

10

Feb 26, 2026

UNOBench real (Easy)

76.2SR (Precision)

10

Feb 26, 2026

UNOBench real No obstructions

72.5SR (%)

10

Feb 26, 2026