Why has the context size been reduced to only 128k?
#4
by ChristopherKai - opened
Splash supports up to 256K context; the default is selected automatically based on the available memory budget. With enough memory, you can request it with splash serve --model incoai/Qwen3.8-27B-Splash --max-context 256K. Which client/version and launch settings are showing the 128K limit?