Unverified50% confidenceFactExact time
GLM 5.2在Frontier SW基准测试中得分为74.4%，接近Anthropic Opus 4.8的水平
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/2/2026
First Seen
Valid until: 9/30/2026
Sources
GLM 5.2深度实测：开源模型硬刚闭源巨头的真实表现
bilibili纳兰仙素6/19/2026
Related Claims
UnverifiedGLM 5.2 scored 74.4% on the Frontier SW benchmark, closely approaching Anthropic's Opus 4.8.83% similarUnverifiedGLM4在代码推理和通用知识测试中得分74.4%，仅次于Claude 3 Opus79% similarUnverifiedGLM 5.2的Diff测试分数从GLM 5.1的18分跃升至46.2分75% similarUnverifiedGLM 5.2在Terminal Bench测试中得分81分，GLM 5.1得分63.5分，GPT 5.5得分84分，Opus 4.8得分85分75% similarUnverifiedGLM4在SWE-Bench基准测试中得分46.2%，显著优于上一代模型74% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/61237API
curl https://kongchang.com/api/v1/knowledge/claims/61237MCP
get_claim(id=61237)