DeepResearch
Benchmarks
TaskDatasetSOTA resultTrendResults
DeepResearch Bench
58.03RACE Overall
58
DeepResearch 1319 queries (test)
74.5Agreement Accuracy
19
DeepResearch Bench (test)
48.24Overall Score
13
DeepResearch
44.7HotpotQA Score
12
DeepResearch benchmark
53.4Average Score
8
DeepResearch Bench
3.72DLB
7
DeepResearch-9K 29 (test)
0.45P-hat
6
DeepResearch
44.21Token Reduction
2
DeepResearch-9K
47.8Omega
1