benchgap
Other · model

Inkling (xhigh) benchmark scores

As of 2026-10-07, Inkling (xhigh) (Other) has measured scores on 13 benchmarks and estimated scores on 1 more.

BenchmarkScoreSource
Terminal-Bench 2.155.1%measured
Terminal-Bench 4.01.0%measured
Terminal-Bench Science 0.10.0%measured
GPQA Diamond87.2%measured
SciCode47.0%measured
CritPt5.4%measured
GDP.pdf12.8%measured
τ³-Bench Banking29.1%measured
MMMU-Pro73.0%measured
AA-AnalystAgent23.8%measured
Harvey LAB91.9%estimated ± 1.2 pp, medium confidence
Humanity's Last Exam31.9%measured
AutomationBench5.0%measured
EnterpriseOps-Gym38.0%measured