Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

MemoryBench

Benchmarks

Task NameDataset NameSOTA ResultTrend
MemoryBench TaskMemoryBench Long-Input-Long-Output
Norm-Score64.23
24
MemoryBench TaskMemoryBench Short-Input-Short-Output
Norm-Score76.26
24
MemoryBench TaskMemoryBench Short-Input-Long-Output 1.0
Norm-Score77.09
24
MemoryBench TaskMemoryBench Long-Input-Short-Output
Norm-Score52.6
22
Robot ManipulationMemoryBench (test)
Success Rate100
7
Robot ManipulationMemoryBench extended
Put Block Back93
4
3D Robotic ManipulationMemoryBench
Avg Success Rate94.3
3
Showing 7 of 7 rows