List your company

Language Understanding on MMLU High tier (Reward and Tool Cost)

77.4Reward+RL Free
71.57673.08874.676.112Jun 5, 2026

Evaluation results

MethodLinks
+RL FreeBackbone=Qwen3-4B, RL strategy=Free2026.0677.472
+RL GateBackbone=Qwen3-4B, RL strategy=Gate, Protocol=confidence-gated GSPO2026.0674.523
+RL OTCBackbone=Qwen3-4B, RL strategy=OTC2026.0673.80
+RL FlatBackbone=Qwen3-4B, RL strategy=Flat2026.06730
BaselineBackbone=Qwen3-4B2026.0671.80