Adobe is giving its Firefly AI video-generation app a major upgrade, aimed at making video creation and correction far less frustrating. The biggest change is a new built-in video editor designed for precise, prompt-based edits—so you can tweak what you already generated instead of starting over from scratch.
Until now, Firefly mainly handled text-to-video generation, which meant that if one part of a clip didn’t look right, the common workaround was to regenerate the entire video and hope for a better result. With the new editor, Firefly users can make targeted adjustments using simple text instructions. You can refine specific visual elements, shift colors, and even alter camera angles. Adobe is also introducing a timeline view, making it easier to manage details like frames, audio, and other clip characteristics in a more familiar editing-style workflow.
Adobe first revealed this editor in October through a private beta, and it’s now rolling out more broadly to all users.
A key part of this update is deeper support for additional AI models inside Firefly. Using Runway’s Aleph model, users can issue more specific edit commands such as “Change the sky to overcast and lower the contrast” or “Zoom in slightly on the main subject.” These kinds of prompt-driven edits are meant to help creators quickly polish scenes without repeatedly regenerating content.
Adobe is also expanding what its own Firefly Video model can do. Users can upload a starting frame and provide a reference video showing the desired camera movement, then instruct Firefly to recreate that camera motion for the clip they’re developing. This should be especially useful for creators who want more consistent cinematic movement across shots or need a particular style of pan, zoom, or motion match.
On the quality side, Firefly is adding support for Topaz Labs’ Astra model, which can upscale video output to 1080p or even 4K—an important feature for anyone trying to turn AI-generated clips into something that holds up on modern displays, presentations, or client deliverables.
For image generation, Adobe is also bringing in Black Forest Labs’ FLUX.2 model. Adobe says FLUX.2 will be available across Firefly platforms immediately, and that Adobe Express users will be able to access FLUX.2 starting in January. Alongside the model expansions, Adobe is also adding a collaborative boards feature, signaling a push toward shared creative workflows where teams can review, iterate, and build on ideas together.
With competition in AI image generation and AI video generation moving quickly, Adobe is clearly trying to make Firefly more compelling—not just as a generator, but as an end-to-end creative environment for editing and refinement. To encourage more users to experiment with these tools, Adobe says subscribers on Firefly Pro, Firefly Premium, and the 7,000-credit and 50,000-credit plans will get unlimited generations from all image models and the Adobe Firefly Video model inside the Firefly app until January 15.
This update continues a busy year for Adobe Firefly. Adobe introduced a subscription option in February covering multiple tiers of image and video generation, then followed up with a refreshed Firefly web app and mobile apps later in the year. The company has also continued adding third-party model support inside Firefly, positioning the app as a place where creators can choose the best model for a task—then edit results with much finer control than before.
