BREAKING
Zhipu AI Releases GLM-5.2 (MIT)
0
B
total params
0
B
active params
0
M
context tokens
Benchmark Results
Terminal-Bench 2.1
81
SWE-bench Pro
62.1
GLM-5.1 SWE
58.4
0
$/M
input tokens
0
$/M
output tokens
Strengths vs Concerns
Strengths
●
Finds bugs others miss
●
Strong on large repos
●
Local, private execution
Concerns
●
Guardrail removal risk
●
Mixes in Chinese text
●
Benchmarks need checking
Open Source Narrows the Gap
AI NEWS BLITZ
China's Zhipu AI has released GLM-5.2 as an open-weight model under the MIT license.