Tag
Preliminary results from Agent Arena Code show good performance for GLM and Qwen models, indicating advancements in open-weight AI models.