Ling-3.0 7.9B (1.3B active) Tiny
on 680M · llama.cpp
Summary
User reports Ling 3.0 Tiny at 15 t/s on an AMD 680M iGPU with llama.cpp, in power saver mode.
Gemma 4 E4B ran at 9 t/s on the same iGPU, also in power saver mode; the laptop stayed cold and silent.
User calls Ling usable and asks whether anyone has run Gemma 4 26B-A4B on a 780M laptop, guessing around 20 t/s by extrapolation.