Adobe Firefly now supports more than ten external AI models—including OpenAI, Google Gemini, Runway and Topaz Labs—allowing creators to generate and refine images, video, audio and text without leaving the platform. The update rolls out to all Firefly users in October 2026.
Visit the official Adobe Firefly website →
What changed
Adobe has added full compatibility with a suite of leading AI models. The list includes Google’s Gemini 2.5 Flash Image (Nano Banana), Imagen, OpenAI’s GPT image generator, Black Forest Labs’ Flux 1.1 and Flux 1 Kontext, Runway Gen‑4, Topaz Labs’ Astra and Bloom, Ideogram, ElevenLabs and Kling AI. Together with Adobe’s own Image Model 5, these integrations let users switch between models for specific tasks—such as photorealistic rendering, precise text‑driven edits, high‑resolution output or voice‑over generation—while staying inside Firefly’s interface.
Who gets it and when
The compatibility expansion is available to all existing Adobe Firefly subscribers as of the October 2026 release. New users who start a free trial or subscribe to a paid tier will see the partner‑model options immediately in the UI. Because the feature is built into the core platform, no additional add‑ons or separate licenses are required.
How to use it
- Open Firefly and choose the asset type you want to create (image, video, audio, etc.).
- In the model selector, pick a partner model or stick with Adobe’s Image Model 5.
- Enter your prompt. If you’re refining an output from another AI (for example, a GPT‑generated image), select the corresponding partner model to apply style variations or enhancements.
- Use Firefly Boards or Foundry to compare results from multiple models side‑by‑side.
- Export the final asset with Adobe Content Credentials attached for provenance tracking.
What it means if you are choosing Adobe Firefly
The expanded model library makes Firefly a strong hub for teams that already juggle several AI tools. You gain flexibility to pick the best‑performing model for each creative need without switching apps, and you keep the commercial‑safe guarantees Adobe promises—ownership of outputs, no training on your content, and built‑in metadata.
However, the breadth of options can add complexity. Users who prefer a single, highly specialized model may find the UI a bit crowded, and licensing terms for third‑party models can differ, so you’ll need to verify commercial rights for each partner. If you primarily need only image generation, dedicated tools like Midjourney or Stable Diffusion‑based services might be simpler and cheaper.
FAQ
Can I use the partner models for commercial projects?
Firefly’s own outputs are commercially safe, but each partner model has its own usage terms. Adobe provides links to those terms, so you should review them before using the assets commercially.
Do I need to pay extra for the new models?
The compatibility feature is included in all Firefly plans as of the October 2026 update. Some partner models may have their own subscription fees, which would be billed separately by the model provider.
Is my data still private with these integrations?
Adobe states that user‑generated content is not used to train Firefly or its partner models. Privacy policies for each third‑party model still apply, so check those if you handle sensitive material.
Bottom line
Adobe Firefly’s new multi‑model support gives creators a unified workspace for a wide range of generative tasks, while preserving Adobe’s commercial‑safe stance. It’s ideal for teams that need flexibility across image, video, audio and text AI tools, but solo creators or budget‑focused users might prefer a single‑purpose platform.
Sources
Full guide: Adobe Firefly Pricing and Free Credits: Plans, Limits, and Commercial Use — plans, prices and alternatives.