Qwen3.8 flash at 150–177 tok/s gen and ~1,900 tok/s prefill on 2x 3090s + 96gb ddr5
bone engine, my own inference engine. 256k context, single stream or 4x concurrent (50ts a stream, up to 256k)
vs strata engine on the same rig: 120-140gen, 1k pp → 150-170gen, 2k pp
iq3_xxs quant, so not full precision, will optimize q4 once qwen 4 drops. Model is leftover from when i had less ram.
I dont want to post a let's connect post but I genuinely want to follow more people in the ai/local space. My feeds too small.
If thats you or you know someone, shout below.