来自MSN

GLM-4.6 首发实测：和 Claude 4.5 比怎么样？

GLM-4.6 昨夜低调放号，我们第一时间拉来 Claude 4.5 做 48 小时盲测。结果出乎意料：中文指令遵循率 GLM 领先 9.4%，代码一次性可运行率反超 7%，更在 2024 高考数学卷拿下 142 分，比 Claude 高 18 分；但在多轮逻辑推理和长程上下文回忆上，Claude 依旧守住“最像人”的 ...
VentureBeat

Chinese startup Z.ai launches powerful open source GLM-4.5 model family with PowerPoint ...

Want smarter insights in your inbox? Sign up for our weekly newsletters to get only what matters to enterprise AI, data, and security leaders. Subscribe Now Another week in the summer of 2025 has ...