Adobe Partner Models | More AI Choices Across Creative Cloud
Adobe is expanding the selection of partner generative AI models available across Creative Cloud. Creators can now choose models from companies including Google, OpenAI, Black Forest Labs, Runway, Luma AI, Topaz Labs, ElevenLabs, and Kuaishou within Adobe Firefly, Photoshop, Illustrator, and Adobe Express.
Adobe brings more partner AI models into Creative Cloud
Adobe has detailed the growing selection of partner generative AI models available inside its creative products. These models support image generation, image editing, video generation, video editing, vector creation, speech generation, and image enhancement across several Adobe applications.
The integrations give creators more choice without requiring them to leave their existing Adobe workspace. Depending on the feature and application, users can select between Adobe's proprietary Firefly models and supported partner models through the model selection menu.
Google models support image, vector, and video creation
Adobe products support several Google models, including Gemini image models and Veo. Gemini models can be available for tasks such as text-to-image generation, prompt-based image editing, Generative Fill in Photoshop, vector creation in Illustrator, Firefly Boards, Firefly Graph, and Adobe Express.
Google Veo supports text-to-video and image-to-video generation in Firefly, along with Firefly Boards, Firefly Graph, and the Firefly video editor. Adobe describes Veo as an option for creators seeking realistic motion and high-quality video output.
OpenAI models expand image and video options
Adobe also offers several OpenAI models. Its supported image models include GPT Image, GPT Image 1.5, and GPT Image 2. Availability varies by model, but supported uses include text-to-image generation, prompt-based editing, Firefly Boards, Firefly Graph, and vector features in Illustrator.
Sora 2 is available for video-related features such as text-to-video generation, Firefly Boards, and the Firefly video editor. Adobe also uses OpenAI language models for specific functions, including GPT-5.1 in Firefly Graph and GPT-4.1 mini for text rewriting in Illustrator.
Runway, Luma AI, and Kling strengthen video production
Adobe's partner selection includes several models focused on video. Runway Gen-4.5 supports text-to-video and image-to-video generation, Firefly Boards, Firefly Graph, and the Firefly video editor. Runway Aleph models provide natural-language video editing and frame-based transformations.
Luma AI Ray models offer video generation and editing with an emphasis on visual quality and realistic motion. Kling models add single-shot and multi-shot video creation, including options for consistent characters, complex scenes, audio, and advanced prompt control.
Photoshop gains more generation and enhancement tools
Partner models are also becoming part of familiar Photoshop features. Depending on current availability, creators can use selected Gemini and FLUX models with Generative Fill, while Topaz models support Generative Upscale, sharpening, and noise reduction.
This gives Photoshop users different tools for different stages of an image. One model may be better suited to generating or editing visual content, while another may be more useful for increasing resolution, recovering detail, sharpening a photograph, or reducing visible grain.
Illustrator and Express receive partner model support
Adobe Illustrator supports partner models in features such as Text to Vector Graphic, Generative Shape Fill, Prompt to Edit, and Rewrite. This extends partner model selection beyond raster image generation and into vector design and text refinement.
Adobe Express also supports selected partner models for accessible content creation. However, Adobe notes that availability depends on the plan. Partner models are not currently included with Adobe Express free accounts, Education plans, or Adobe Express for Enterprise.
ElevenLabs brings multilingual speech into Firefly
ElevenLabs Multilingual v2 is available for Generate Speech in Firefly and Text to Speech in the Firefly AI Assistant beta. Adobe describes the model as capable of producing natural speech with emotional range across multiple languages.
This broadens Adobe's partner model strategy beyond visual generation. Creators can combine image, video, vector, editing, enhancement, and speech capabilities inside a connected Adobe production environment.
Partner models use Adobe generative credits
Each partner model can consume a different number of Adobe generative credits. Image generation may use a fixed number of credits based on the selected model and resolution, while video models commonly consume credits for every generated second.
Adobe warns that consumption rates, plan access, model availability, and promotional pricing can change. Users should review the current credit rate before running high-resolution images, long videos, repeated variations, or large production batches.
Adobe explains how prompts and reference files are handled
Adobe says content submitted through its products is not used to train partner generative AI models or Adobe models. The company only offers partner models when the developer or hosting platform contractually agrees to that condition.
When a partner model is used, Adobe shares the prompt and any uploaded reference image or video required to generate the result. Adobe also attaches Content Credentials to images created entirely with generative AI, helping users identify whether AI was used and which model generated the content.
IMPORTANT: Partner models are not developed by Adobe. Creators are responsible for deciding whether a model is appropriate for a project. Availability, credit consumption, plan requirements, data handling, and supported features can vary by model and may change over time.{alertWarning}
Daisuki's Take: What This Means for Designers
Adobe's partner model strategy gives designers more freedom to choose the right model for each creative task without constantly moving files between unrelated platforms. Image generation, video production, vector creation, speech, and enhancement can now connect more naturally with familiar Adobe tools.
The important part is that these models are not interchangeable. A designer may prefer one model for prompt accuracy, another for typography, another for cinematic motion, and another for upscaling. Understanding those differences will be more valuable than always choosing the newest option.
We would also watch credit usage carefully. More model choice can improve creative control, but testing many high-resolution generations or video variations can consume credits quickly. The best approach is to match the model to the task, test with smaller outputs when possible, and reserve expensive generations for directions that are already working.
Sources and Recommended Links
- Partner models in Adobe products | Adobe Official Help Center
