qwen-3.8-max — GPT