List your company

User Simulation Quality Assessment on Session-level Human Evaluation Set adversarial (test)

86Win RateUserLM-R1
42.3253.666576.34Jan 14, 2026

Evaluation results

MethodLinks
UserLM-R1Opponent=Gemini-2.5-Flash2026.01862212
UserLM-R1Opponent=UserLM-R1 (w/o RL)2026.01553233
UserLM-R1Opponent=DeepSeek-R12026.01442848