community pool / model answer

Phi-4-reasoning

chat · microsoft/Phi-4-reasoning

31.4 tok/s · model median · 26 runs in the model record.

Reference rigQuantizationMedian resultBasisTTFT (ms)Peak VRAMMax contextEngines
RTX 3060 12GB ×24-bit28.1 tok/smeasured2,47617.2 GB131,072llama.cpp
RTX 3090 24GB4-bit41.2 tok/smeasured-- GB131,072llama.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.