BREAKING
Z.ai Releases Open-Weight GLM-5.2
0
B
Total params
0
B
Active per token
0
M
Context tokens
SWE-bench Pro Scores
GLM-5.2
62.1
GLM-5.1
58.4
GPT-5.5
58.6
Opus 4.8
69.2
0
$
GLM-5.2 per task
0
$
GPT-5.5 per task
Strengths and Open Questions
Strengths
●
Repo-wide coding
●
Agentic bug fixes
●
First open daily driver
Concerns
●
No image input
●
Inference speed limits
●
Trails Opus at times
Open Alternative for Developers
AI NEWS BLITZ
China's Z.ai has released its flagship GLM-5.2 model as open weights under the MIT license.