Thread by @AskVenice

Introducing Venice Image Engine v2 🌅

Venice Image Engine v2 is now live and represents a comprehensive overhaul of our image generation infrastructure, delivering the highest quality results across a range of styles

Let’s dive in 🧵

Venice Image Engine v2 consists of two major components:

1. Venice SD35 (New Default model): Custom-configured Stable Diffusion 3.5 engine powered by a Comfy UI workflow backend.

2. Upscale & Enhance (Pro only): Completely new upscaling architecture with proprietary enhancement that creatively in-fills new pixels at the higher resolution.

Venice SD35

Venice implemented substantial improvements to the base Stable Diffusion 3.5 architecture:

Advanced prompt processing: We've removed token limitations that restrict other image generators, allowing you to write longer, more detailed prompts that better capture your creative vision.

Natural language understanding: Venice SD35 processes conversational language more effectively, so you can describe images as you would to a human artist instead of using awkward keyword lists.

Custom image generation pipeline: a specialized image generation Comfy UI workflow on the backend that delivers superior results through optimized processing techniques that standard implementations don't provide.

Let’s compare our new default image model Venice SD35 to our previous default image model Fluently

Upscale and Enhance

The new Upscale & Enhance system represents a significant leap beyond traditional pixel-by-pixel methods:

Use Upscale for accurate enlargements: Use the Upscale feature when you want to maintain the exact style and composition of your image while increasing resolution.

Use Enhancer for creative enhancements: Use the Enhance feature when you want to add details and refinements beyond what's in the original image. Adjust the creativity setting based on how much creative liberty you want the AI to take. For consistent results we suggest loading in your original prompt in the prompt field.

Combine for professional outputs: Combine both approaches by first generating with Venice SD35, then upscaling, and finally applying a subtle enhancement for the highest quality results.
Let’s compare zoomed-in screenshots for the following image.

Let’s compare zoomed-in screenshots for the following image.

Venice SD35 standard

vs

Venice SD35 + 4x Upscale

vs

Venice SD35 + 4x Upscale + Enhance

For more details, as well as a comprehensive guide on best prompting tips to get the most out of Venice Image Engine V2 + prompt template, read our blog:

venice.ai/blog/introduci…

You can try the new Venice SD35 model for free at

Share your best images with usvenice.ai