List your company

Web-based QA on BrowseComp+

10AccuracyOrch-RM
2.85524.71016.5658.4199Jun 11, 2026

Evaluation results

MethodLinks
Orch-RMContext=Continued-training, Sampling Level=Orchestration2026.061011
LLM-as-a-judgeContext=Continued-training, Sampling Level=Orchestration2026.069.639
DPOContext=Continued-training2026.069.569
MAS-OrchestraContext=Training-from-scratch, Sampling Level=Trajectory2026.069.449.5
RFTContext=Continued-training, Sampling Level=Trajectory2026.067.316
GRPOContext=Continued-training, Sampling Level=Trajectory2026.066.946
Orch-RMContext=Training-from-scratch, Sampling Level=Orchestration2026.065.944.5
Qwen2.5-7B-InstructContext=Base orchestrator2026.063.130.5