community pool / model answer

Llama-4-Scout-17B-16E-Instruct-GGUF

chat · unsloth/Llama-4-Scout-17B-16E-Instruct-GGUF

44.9 tok/s · model median · 2 runs in the model record.

Answer

44.9tok/s

reportedCMP 170HX 64GB3-bit · 2 runs · decodebest cell in the pool

  1. CMP 170HX 64GB44.9 tok/sreported2 runs · peak 46.4 GB · fits 64 GB rig memory
Reference rigQuantizationMedian resultBasisTTFT (ms)Peak VRAMMax contextEngines
CMP 170HX 64GB3-bit44.9 tok/sreported-46.4 GB2,048llama.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.