guys, I ma testing now the @matthew16 's GitHub - kindlingai/glm-5.3-flash-gx10: GLM-5.3-Flash (NVFP4) at TP=4 across 4x ASUS GX10 / GB10 - vLLM + mentat instead of Ray, 524k context, DFlash2. Minimal working recipe. · GitHub for TP2 and this is the next ceiling for us.
BTW. Have you saw him GitHub - kindlingai/kindling-spark-os: Minimal read-only OS image for GB10 (DGX Spark / ASUS GX10) boxes · GitHub ? The most valuable company in the world couldn’t do what one person did.
Here is the post in our community Kindling Spark OS - ~4GB+ free RAM