RealDataAgentBench

Most LLMs get the right answer. RDAB checks if they did it the right way.

— runs — tasks — — GitHub ↗ Scoring Spec ↗
# Model RDAB Score Coverage Avg Cost / Task Provider
Task ↕ Title ↕ Diff. ↕ Category ↕ Model ↕ Runs ↕ Correctness ↕ Code Quality ↕ Efficiency ↕ Stat Validity ↕ RDAB Score ↓ Tokens ↕ Avg Cost ($) ↕ Score / $