Firefly's Latest Update Allows Users to Combine Adobe, OpenAI, and Google AI Models in One App: Is Creative Safety Still Guaranteed?

Adobe Expands Firefly With OpenAI And Google AI Models For Mobile Creativity

Announced on Thursday, Adobe is giving creators more tools than ever by bringing top-tier AI image and video generation models into its Firefly platform, including models from OpenAI and Google.

With this move, the company is combining its proprietary AI capabilities with some of the most advanced third-party models available, offering users the freedom to choose how they create — all while reinforcing its promise of legal safety for content produced with its own models.

Why Adobe Is Integrating External AI Models Into Firefly

Since 2023, Adobe has focused on developing its own Firefly AI models, specifically designed to generate images and videos that are legally safe for commercial use.

Now, Adobe is embracing flexibility by allowing external models into the Firefly environment.

These include OpenAI’s image generation model from GPT, Google’s Imagen 3, Veo 2, and Flux 1.1 Pro.

Ely Greenfield, Adobe’s CTO for digital media, explained this dual approach during an interview with Reuters on Monday,

"We still have lots and lots of customers for whom taking stuff to production, they will only use Firefly because the commercial safety really matters to them. But for other parts of the workflow, like ideation, they're interested in experimenting with other models as well. So we're making that choice available to them."

What Makes Firefly Image Model 4 And 4 Ultra Stand Out

Among Adobe’s latest proprietary offerings is the Firefly Image Model 4, which the company calls its most realistic and controllable model to date.

It allows for up to 2K resolution output and gives users detailed control over camera angles, zoom, and visual depth.

The “Ultra” version of this model further enhances realism and excels in producing complex scenes with intricate details.

Firefly Video Model And Text-To-Vector Tools Now Available

Adobe has also launched its Firefly Video Model to general availability, moving it beyond beta testing.

This tool enables users to create high-quality video clips from text prompts or still images, with creative controls like cinematic camera motion.

Additionally, the new Firefly Vector Model now powers Adobe’s Text-to-Vector feature, giving designers more options for creating scalable illustrations directly from words.

How Adobe Is Keeping Content Legally Safe And Transparent

One of Adobe’s biggest advantages remains the commercial safety of its Firefly-generated content.

Unlike some third-party models, Adobe guarantees that anything created with Firefly is free from intellectual property risks.

To ensure transparency, all content created using any model within Firefly now includes Adobe’s Content Credentials — a digital “nutrition label” that displays information such as creator details and the tools used in the process.

What Sets OpenAI’s GPT-4o Apart From Other Models

The newly included OpenAI GPT-4o model has a standout feature that may attract users for specific tasks: it can generate accurate, high-quality text within images — a common struggle for most models, including Adobe’s own.

In workflows where text precision in images is key, being able to switch to OpenAI’s model offers a practical advantage.

Enterprise Controls And How Model Choice Works In Firefly

Adobe makes it clear which model is being used at any given time, providing peace of mind for users concerned about legal safety or specific output capabilities.

For enterprise users, there is the added ability to enable or restrict the use of third-party models within their organisations.

All models — including Firefly Image Model 4, Image Model 4 Ultra, and the Firefly Video Model — are now live on Firefly via web.

The mobile version of the app is on its way for both iOS and Android users, with pre-orders currently open.

Adobe’s Firefly Boards feature is also available in public beta on the app.

More Partner Models On The Way To Firefly

Adobe plans to add more AI models from external partners including fal.ai, Ideogram, Luma, Pika, and Runway.

All will be accessible within the Firefly interface, and users can pay for them using the same credit system they already use for Adobe’s tools — though the company has not disclosed how revenue will be shared between Adobe and its AI partners.

With over 22 billion assets generated through Firefly since launch, Adobe is aiming to give its creative community more power and more choices — wherever they choose to work.