List your company

Agentic Performance Aggregate on Combined LCtx, ALFWorld, ScienceWorld, Web/Tool

78.9Mean Task SuccessCVT-RL
45.93254.49163.0571.609Jun 3, 2026

Evaluation results

MethodLinks
CVT-RLVerifier=yes, Constraints=yes, CF rollouts=semantic, Belief=yes, Compute=3.92026.0678.982.83.90.175
Information-matched CF processVerifier=yes, Constraints=yes, CF rollouts=semantic, Belief=yes, Compute=3.92026.0675.478.97.20.224
Compute-matched non-causalVerifier=yes, Constraints=yes, CF rollouts=random extra, Belief=yes, Compute=3.82026.0671.874.611.70.271
Constrained process RLVerifier=yes, Constraints=yes, CF rollouts=no, Belief=yes, Compute=1.52026.0669.573.812.60.286
RLVMRVerifier=yes, Constraints=no, CF rollouts=no, Belief=no, Compute=1.32026.0668.772.913.50.305
LongRLVRVerifier=yes, Constraints=no, CF rollouts=no, Belief=no, Compute=1.22026.0667.170.814.80.339
Q-RAGVerifier=yes, Constraints=no, CF rollouts=no, Belief=no, Compute=1.42026.0666.873.514.10.332
TROLL-styleVerifier=yes, Constraints=no, CF rollouts=no, Belief=no, Compute=1.12026.0663.566.415.90.351
PPO-RLVRVerifier=yes, Constraints=no, CF rollouts=no, Belief=no, Compute=1.12026.0658.463.117.40.384
SFTVerifier=no, Constraints=no, CF rollouts=no, Belief=no, Compute=1.02026.0647.255.818.90.412