# Stable Diffusion 3.5: What's New in the Latest Generation

DevFeed: [Stable Diffusion 3.5: What's New in the Latest Generation](<https://devfeed.tech/articles/stable-diffusion-3-5-what-s-new-in-the-latest-generation-94589.md>)

Original publisher: [Read original article](<https://www.runpod.io/blog/stable-diffusion-3-5-update>)

Author: Brendan McKeag

Published: 2026-10-22T02:56:23Z

Content type: article

Language: en

Sources: [Runpod Blog.](<https://devfeed.tech/sources/runpod-blog.md>)

Topics: [stable-diffusion](<https://devfeed.tech/topics/stable-diffusion.md>), [ai-image-generator](<https://devfeed.tech/topics/ai-image-generator.md>), [Enhancing AI model performance](<https://devfeed.tech/topics/enhancing-ai-model-performance.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [ai-art](<https://devfeed.tech/tags/ai-art.md>), [animals](<https://devfeed.tech/tags/animals.md>), [community](<https://devfeed.tech/tags/community.md>), [cuda](<https://devfeed.tech/tags/cuda.md>), [discord](<https://devfeed.tech/tags/discord.md>), [docker](<https://devfeed.tech/tags/docker.md>), [docker-image](<https://devfeed.tech/tags/docker-image.md>), [efficiency](<https://devfeed.tech/tags/efficiency.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [generate](<https://devfeed.tech/tags/generate.md>), [gpu](<https://devfeed.tech/tags/gpu.md>), [image](<https://devfeed.tech/tags/image.md>), [image-generation](<https://devfeed.tech/tags/image-generation.md>), [images](<https://devfeed.tech/tags/images.md>), [models](<https://devfeed.tech/tags/models.md>), [quality](<https://devfeed.tech/tags/quality.md>), [sd3](<https://devfeed.tech/tags/sd3.md>), [sd3-5](<https://devfeed.tech/tags/sd3-5.md>), [stable-diffusion](<https://devfeed.tech/tags/stable-diffusion.md>)

## AI overview

Stable Diffusion 3.5 launched in Large and Large Turbo variants: Large prioritizes image quality, while Turbo favors efficiency; a Medium model for smaller GPUs was expected the following week. The excerpt reports improved anatomy relative to SD3 and photorealistic people and animals from minimal prompts, with images up to 2 megapixels. Results are not uniformly reliable: a prompt for a black-and-white Argentine tegu produced inaccurate reptiles. For quick experimentation, the author says the largest model can run on a Runpod pod in under two minutes for well under $1 per hour. A Camenduru serverless image is also available; the article says users can clone the repository and build their own Docker image, while a Runpod Quick Deploy serverless image was still forthcoming. The model was newly released, so community checkpoints and fine-tunes were not yet established.

## Source excerpt

Stability.ai's SD3.5 is here, with new models built for speed and quality. Learn what's changed, what's improved, and how to run it on Runpod.