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

Lifelong Agent Bench

Benchmarks

Task NameDataset NameSOTA ResultTrend
OS TaskLifelong Agent Bench OS Task
Success Rate (Last Epoch)78.8
11
DB TaskLifelong Agent Bench DB Task
Last Epoch Success Rate96
6
DB TaskLifelong Agent Bench (val)
Success Rate94.2
6
OS TaskLifelong Agent Bench (val)
Success Rate74.6
6
Showing 4 of 4 rows