BREAKING
GLM-5.2 tops open-weight models
0
Intelligence Index
0
Points over GLM-5.1
0
Overall rank
Context window jumps to 1M
GLM-5.1200
GLM-5.21000
0B
Total params
0
SWE-bench Pro
0
GPQA Diamond
Cheap, open, long-context play
GLM-5.2Open weights
●MIT license on Hugging Face
●$1.40 in / $4.40 out per 1M
●Plan ~1/10 of Claude-class
Tradeoffs
●Tight rate limits on cheap tiers
●Frontend-skewed evals
●No benchmarks at launch
Open-weight race heats up
AI NEWS BLITZ
Z.ai has released GLM-5.2, now the highest-scoring open-weight model yet.