Adobe is closing out 2025 with a fresh slate of AI upgrades designed to make video creation and editing easier for creators. On Tuesday, the company is upgrading the availability of its Firefly video editor, along with announcing new tools for precise edits and camera-motion control, a video upscaling feature in Firefly Boards, and support for the FLUX.2 model.
As a holiday bonus, Adobe is also giving Firefly subscribers unlimited generations through January 15.
“Our goal with Adobe Firefly is to give you one place where you can choose the right industry-leading model for your project, then get exactly what you want out of it by using it with the creative tools and controls Firefly offers,” Mike Polner, Adobe’s vice president of product marketing and its general manager for creators, writes in a blog post.
Adobe has been on a tear as of late, turning Firefly into a comprehensive multimodal app for creators. It supports not only text-to-image but also text-to-video and text-to-audio. Today’s investment aims to provide users with the controls needed to fine-tune their video generations to make them appear as professional-looking as possible—almost as if a human had done it.
Subscribe to The AI Economy
Status Update for Firefly Video Editor
The first announcement from Adobe involves its Firefly video editor. Unveiled initially at Adobe Max earlier this year, this app is now open to the public in beta—it was previously by invite-only. As a refresher, the Firefly video editor is a browser-based app built to help creators turn their video generations into polished stories. The company claims that entire videos can be created with this application. It also supports blending generated clips, music tracks, and visuals with a creator’s own footage and then producing them on a “lightweight” multi-track timeline.

There are two ways to edit a video in the editor: a traditional timeline view, which offers granular control over pacing, cuts, and layers, or a text-based workflow that’s especially well-suited for interviews, letting creators trim or rearrange segments simply by editing the text in the transcript.
When a video is ready for distribution, it can be exported in multiple formats, including vertical social posts to widescreen edits.
Is this the version of Premiere that’s needed for the AI era? The goal is to provide something with the power, but with less complexity than Adobe’s video app. That said, the company isn’t the only one with such a tool—Runway and DaVinci Resolve are a couple of the software makers with similar offerings.
“We know that generating clips, downloading them, and using them in Premiere is a workflow customers do today,” Alexandru Costin, Adobe’s vice president of generative AI and Sensei, said in October. “We want to make this workflow even easier. The Firefly video editor…enables…nonlinear editing, a timeline-based editing generation, importing videos, applying effects, and bringing all of that together into a really functional web video editor.”
New Firefly Editing Tools
Next up are two new Firefly video editing tools that eliminate the need for regenerating entire clips if a mistake is found.

With “Prompt to Edit” controls, creators can make precision refinements to Firefly-generated videos using text editing. Using natural-language prompts such as “remove the person on the left side of the frame” or “change the sky to overcast and lower the contrast,” this capability enables surgical-like editing, powered by Runway’s Aleph model. All changes are made directly to the existing clip.

The second new tool, called “camera motion reference,” gives creators control over how the camera moves through a scene. Using the Firefly Video Model, users upload a start-frame image along with a reference video showing how they want the camera to move. Then, like magic, Firefly will generate “cinematic camera movements” anchored exactly where the creator wants them.
Both of these capabilities promise to save hours of trial and error, along with money, since creators no longer have to regenerate entire scenes and spend precious credits. If there was a message Adobe wants to send, it’s perhaps this: Creators are the directors, not AI.

But those aren’t the only tools from Adobe. There’s still one more, and it’s specific to Firefly Boards: Video upscaling. Powered by Topaz Astra, an AI-powered video enhancement offering from Topaz Labs, creators can upscale footage to 1080p or 4K.
Do you have a low-resolution clip and want to make it optimized for YouTube? Want to restore older or low-quality footage from your archives and make it clear and detailed for today’s audience? Adobe touts that its video upscaling tool can breathe new life into those moments, making them formatted for Instagram, TikTok, Snapchat, and YouTube.
Adding a New Model to Firefly
AI generation in Firefly was never exclusively provided by Adobe’s own models. Instead, the company has embraced the value of choice, adding numerous third-party LLMs to its app. This year has seen a bevy of additions, including those from OpenAI, Google, Ideogram, Pika, Luma, Runway, Moonvalley, and Topaz Labs.
Today, Adobe is announcing the addition of Black Forest Labs’ FLUX.2, a model known for producing photorealistic, high-resolution images. The LLM is available in Firefly’s text-to-image option, Prompt to Edit, Firefly Boards, and also in Photoshop’s Generative Fill on the desktop. The company reveals Adobe Express support will be coming in January 2026.
Unlimited Video Generation
To motivate creators to take advantage of these new capabilities, Adobe is running a promotion through January 15, 2026. Firefly Pro, Firefly Premium, 7,000-credit, and 50,000-credit plan subscribers can now generate unlimited image and video generations through the Firefly web and mobile apps.
It’s not limited to Adobe’s Firefly and Firefly Video models either. Free generations can be made using third-party partner models, including Google’s Nano Banana, FLUX.2, and OpenAI’s GPT Image.
Subscribe to “The AI Economy”
Exploring AI’s impact on business, work, society, and technology.


Leave a Reply
You must be logged in to post a comment.