>>89129530there apparently might be new loras to enhance some things, maybe even longer results but they haven't released all that yet, but they did release this orbit lora. also as of today this is an even better open source model (cogvideox-5b-1.5) but its not ready for inference just yet unless you have a server. waiting on diffuser code but it claims it can i2v at any fucking resolution and the t2v also looks better and can go to 10 seconds so I think shit is about to slap.