Expired50% confidenceFactExact time
GLM-5 scored 77.8% on the SWE-Bench Verified benchmark, placing it in the top tier among open-source models.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/2/2026
First Seen
This claim has expired. It is kept for the record and no longer describes the current state.
Valid until: 9/30/2026(expired)
Sources
GLM-5.2 Deep Dive: Million-Token Context, MIT Open Source & Full-Stack Domestic Chip Training
bilibiliAI技术投降派6/14/2026
Related Claims
UnverifiedGLM4 scored 46.2% on the SWE-Bench benchmark, significantly outperforming the previous generation model.79% similarUnverifiedGLM 5.2 scored 81 points on Terminal Bench, GLM 5.1 scored 63.5 points, GPT 5.5 scored 84 points, and Opus 4.8 scored 85 points.78% similarUnverifiedGLM4 ranks fifth in the World of AI benchmark tool.78% similarUnverifiedGLM 5.2 scored 74.4% on the Frontier SW benchmark, approaching the level of Anthropic Opus 4.8.74% similarUnverifiedGLM 5.2's Diff test score jumped from 18 points in GLM 5.1 to 46.2 points.73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/46081API
curl https://kongchang.com/api/v1/knowledge/claims/46081MCP
get_claim(id=46081)