AI

Z.ai、オープンウェイトのGLM-5.2でGPT-5.5を複数長期的コーディングベンチマークで1/6のコストで凌駕

Z.ai’s open-weights GLM-5.2 beats GPT-5.5 on multiple long-horizon coding benchmarks for 1/6th the cost

venturebeat2026年6月16日

日本語要約

Z.aiの新しいオープンウェイトGLM-5.2モデルが、長期的コーディングタスクにおいてGPT-5.5を上回りました。コストを大幅に削減しながら達成されたこの成果は、強力なAIモデルの急速な進歩とアクセシビリティの向上を浮き彫りにしています。

English Summary

Z.ai's new open-weight GLM-5.2 model has surpassed GPT-5.5 on long-horizon coding tasks. This achievement, at a fraction of the cost, highlights the rapid progress and increasing accessibility of powerful AI models.


元記事を読む𝕏 でシェア
Z.ai、オープンウェイトのGLM-5.2でGPT-5.5を複数長期的コーディングベンチマークで1/6のコストで凌駕 | Mirai Signal