Any update on hosting these models? I keep refreshing the models list to see a new update.
Hi obajohn75,
The build team is constantly trying to bring the latest updates to build.
Stay tuned!
Best,
Aharpster
Could we have an ETA?
appear 07/02/2026 because glm 5.1 is going removed at this date
Is this a pattern youâve seen before? The addition of a new model version right when the old one gets deprecated?
yes allready seen that pattern, also sometime the new model appear just befor the old one is removed, appear that i not the current case we can think that glm5.2 is going appear when glm 5.1 is removed but not 100% guaranted
I hope youâre right. Iâd love to try GLM 5.2
yeap i hope too ^^
edit: appear that they allready removed glm 5.1 and currently not added glm 5.2, kimi k2.6 going deprecated too the 7/7/26 and currently not added glm 5.2 and kimi k2.7 ⌠that is realy strange and diverge from older pattern :/
glm-5.2 | NVIDIA NGC This was just updated on the 30th so maybe its coming soon?
fwiw I can still see it and talk to it in playground here https://build.nvidia.com/z-ai/glm-5.1/playground
itâs not strange at all. when that happened to previous glm versions, we had to wait several days without any glm model before they added the newest model at the time.
Do you remember how many days it took for the last update? and do you remember if the container launch date was at all related to when it was added for us?
It took more or less a week. Hopefully it wonât be any longer. Already missing GLM !
A week!? âšď¸
Itâs already here ! Never been so happy to be wrong :D
Feels pretty quantized though :(
GLM-5.2 supports a 1M-token context window. However, when I use it via the Kilo code I get constant 400 errors when the context grows past 250k tokens. Force limiting context size to 200k makes long agentic tasks much more stable. Is this due to hardware overload?
i have opend issue about that seem api limitation ISSUE GLM 5.2