community pool / model answer

GLM-4.7-Flash-GGUF

chat · unsloth/GLM-4.7-Flash-GGUF

66 tok/s · model median · 9 runs in the model record.

Reference rigQuantizationMedian resultBasisTTFT (ms)Peak VRAMMax contextEngines
Arc Pro B70 32GB4-bit40.8 tok/sreported206- GB4,096llama.cpp
Radeon AI Pro R9700 32GB2-bit82.1 tok/sreported250- GB2,048llama.cpp
Radeon AI Pro R9700 32GB4-bit75.8 tok/sreported196- GB2,048llama.cpp
Ryzen AI Max 395 128GB4-bit52.8 tok/sreported21117.6 GB4,096llama.cpp
Ryzen AI Max 395 64GB4-bit65.7 tok/smeasured1,17517.3 GB8,192llama.cpp

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.