Adobe Releases New Firefly Generative AI Model

Adobe is enhancing its AI-powered creative tool suite by launching the new Firefly generative AI model alongside the newly released Firefly application.

Today, at the company's Max Creativity Conference held in London, Adobe unveiled a brand-new unified Firefly web application for AI-assisted content creation and production. This allows creators to generate images, videos, audio, and vector graphics. The application utilizes Firefly, a series of commercially safe generative AI models developed by Adobe.

To support this application, Adobe has launched two new generative AI models: Firefly Image Model 4 and Firefly Image Model 4 Ultra. The company describes Model 4 as the fastest, most flexible, and most realistic yet, delivering high-quality images with 2K resolution and offering more creative control over camera angles, zoom, and speed than previous versions. The Ultra version enables rendering complex scenes with small structures that require intricate details.

The company also announced that the Firefly Video Model is now fully available. First introduced in the public beta version in February, this model allows users to create 1080p video clips from text prompts or images, modify video camera compositions, produce professional-quality B-roll footage, and generate imaginative elements and custom motion designs based on descriptions.

The company’s new text-to-vector graphics feature is now fully operational, powered by the Adobe Vector Model. This empowers creative designers to produce fully editable vector-based artwork, logos, product packaging designs, icons, patterns, and more, using simple text prompts written in everyday language.

In addition to the new Adobe AI models, the company has begun rolling out third-party generative AI models built by partners, which users can choose to generate their images and videos. The first batch of partners includes OpenAI’s GPT-4o image generation capabilities, Google LLC’s Imagen 3 and Veo 2, and Black Forest Labs Inc.’s Flux 1.1 Pro.

The company stated that it plans to integrate more partner models in the coming months, including those from Fal.ai, Ideogram, Luma, Pika, and Runway.