community pool / model answer

GLM-5.2-MXFP8-NVFP4-NF3-Hybrid

chat · madeby561/GLM-5.2-MXFP8-NVFP4-NF3-Hybrid

74.4 tok/s · model median · 4 runs in the model record.

Reference rigQuantizationMedian resultBasisTTFT (ms)Peak VRAMMax contextEngines
RTX PRO 6000 Blackwell 96GB ×44-bit74.4 tok/sreported175- GB480,000vllm
RTX PRO 6000 Blackwell 96GB ×48-bit68.4 tok/sreported283- GB480,000vllm

Context tested

The table above is the tested context for each exact rig and quantization cell. Empty fields remain empty rather than being filled with an estimate.