List your company

Machine Translation on Challenge EN-ZH

24.44BLEUPEGRL
21.559222.307123.05523.8029Feb 3, 2026

Evaluation results

MethodLinks
PEGRLBackbone=Qwen3-0.6B, RL training=PEGRL, Rollouts per example=722026.0224.4468.8967
MT-R1-ZeroBackbone=Qwen3-0.6B, RL training=MT-R1-Zero, Rollouts per example=722026.022366.8965.28
Qwen3-0.6BTraining=Pre-trained only2026.0221.6764.3363.9