List your company

Task-Focused Dialogue on Multiwoz (G-Eval)

3.453G-Eval ScoreGPT-5.2
3.3493.3763.4033.43Jan 24, 2026

Evaluation results

MethodLinks
GPT-5.22026.013.453
GOPO-Qwen3-14BNumber of parameters=14B2026.013.447
Gemini-2.52026.013.407
GLM-4.72026.013.381
DeepSeek-R12026.013.368
Qwen-235BNumber of parameters=235B2026.013.353