ResearchTasksStep-level reasoning evaluationFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedPDDL (test)Llama-3.1-8B-PRM800k-PDDL-r94.5F1 Score20Apr 21, 2026Rooms (test)Qwen2.5-Math-7B-PRM800k0Error Rate10Apr 21, 2026PRMBench step-level 2025PRISM47.13FPR5Jun 9, 2026