Local video models, tested for real.

text-to-video · Model

LTX 2.5 22B distilled 22B distilled transformer, Comfy int8 convrot (ltx-2.5-22b-distilled-transformer-comfy-int8-convrot), with the int8 Gemma 4 12B text encoder

Run it on a Spark for still-camera product and lifestyle shots at 1920x1088: about three minutes per five-second clip, the subject stays solid, and a single small motion (a head turn, a hand writing, something spilling) comes out clean.

36 s of compute per second of video · Tested Oct 31, 2026