Snap has unveiled an AI text-to-image model for mobile devices that will power some of Snapchat’s features in the coming months. The company said on Tuesday that the model can produce high-resolution images in around 1.4 seconds on an iPhone 16 Pro Max.
The new diffusion model runs entirely on the device, which reduces computational costs compared to models that rely on large servers. Snap notes that the model can create “stunning” visual results by “transferring rich representations from large-scale diffusion models to this smaller and more efficient model.”
Snap plans to bring the technology into production in the coming months to help power the app’s AI Snaps, AI Bitmoji Backgrounds, and more.
The company says that by implementing this in-house technology, it will be able to offer its community high-quality AI tools at a lower operating cost.
“Snap has a long history of research excellence in model optimization and efficiency,” the company said in a blog post. “We are inspired by the industry innovation that is making AI tools more efficient, affordable, and accessible and we look forward to continuing to contribute to the rapid pace of innovation, particularly for mobile-first experiences.”
The company didn’t share any further details about the model but noted that it’s the continuation of its long-term investment in AI and ML technologies.
While Snap has tapped AI tools from companies like OpenAI and Google in the past to power some of its features, it has now debuted its own in-house AI model that will allow it to offer unique features to its users. It makes sense for the company to invest in AI and develop its own model, especially since nearly every tech company, including Meta, has been doing so over the past few years.
Snap is scheduled to release its Q4 2024 earnings report on Tuesday.
