community pool / model answer

Nemotron-Cascade-2-30B-A3B

chat · nvidia/Nemotron-Cascade-2-30B-A3B

155.3 tok/s · model median · 13 runs in the model record.

Reference rigQuantizationMedian resultBasisTTFT (ms)Peak VRAMMax contextEngines
M3 Ultra 512GB4-bit118.1 tok/sreported719- GB8,192llama.cpp, mlx
Radeon AI Pro R9700 32GB ×38-bit89.8 tok/sreported9,971- GB31,620llama.cpp
RTX 3060 20GB4-bit104 tok/sreported-17.6 GB2,048llama.cpp
RTX 3090 24GB4-bit188.1 tok/smeasured1,7690 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.