DIVE: Scaling Diversity in Agentic Task Synthesis for Generalizable Tool Use
Open paper
Benchmark
Model
Rank
Results
agents-on-gaia-text-only-103-validation-tasks
Dive-8B (RL)
#4
Accuracy: 61.2
agents-on-toolathlon
Dive-8B (RL)
#8
Pass@1: 0.083
question-answering-on-humanity-s-last-exam-hle
Dive-8B (RL)
#32
Accuracy: 17.8
Rank counts only results with a code link.