Adobe MAX London is underway today, and the company kicked off the event by announcing the launch of a new generative AI model, Firefly Image Model 4.
Whenever Adobe intros a powerful new Firefly feature it also thoughtfully provides a snapshot of how the model is performing out there in the real world amongst real world users. So far, as things stand in April 2025, over 22 billion assets have been generated with it.
This is impressive. When it first moved out of beta in September 2023 that figure stood at 2 billion. As you can see the usage has accelerated too, and currently stands at something in the order of 1.5 billion assets generated per month.
As well as highlighting the movement of the Firefly Video Model out of beta and into wider availability, Adobe has used MAX to announce details of the newly launched Firefly Image Model Four and its more detailed Ultra cousin.
Essentially, Firefly Image Model Four offers more life-like image quality, more creative control over the structure and style of outputs, and more control over camera angles and zooms. It also accelerates the ability to rapidly explore and iterate concepts and can generate outputs in up to 2k resolution.
The base-spec Image Model 4 is billed as being ideal for rapid ideation and everyday creative needs. Image Model 4 Ultra is pitched at users who are working on projects that demand ‘impeccable detail and realism’, and Adobe says it excels at rendering complex scenes with small structures. Presumably it will use more generative image credits as a result.
Back at MAX in Miami the company showed a sneak of something it called Project Concept, and this is now productised and in beta as Firefly Boards. It's an AI-first surface for moodboarding essentially; an internet-connected creative space designed for brainstorming, iteration, and exploring creative concepts.
It looks pretty powerful too, so we'll have more on that later we hope.
The company has also thrown open the Firefly system to allow ideation with non-Adobe models. This is an increasingly common trend as companies realise the prohibitive costs of creating AI models that cover absolutely all of the bases. The upshot is that while you can keep to the commercially safe (and Adobe cannot emphasise that point too much from the MAX London stage) Firefly models, you can also see content generated by others with different aesthetic styles.
The starts with OpenAI’s image generation capabilities, Google’s Imagen 3 and Veo 2 and Flux 1.1 Pro. Adobe also plans to integrate additional models from partners including fal.ai, Ideogram, Luma, Pika and Runway in the coming months. Creators can easily switch between models, with full transparency into which model is being used at every step.
And lastly, while details are scarce, a Firefly app is coming to mobile soon too.
Pricing and Availability
Firefly Image Model 4, Firefly Image Model 4 Ultra and new Firefly Video Model are all now generally available through Firefly on the web. Firefly Boards is available in public beta in the Firefly app. Firefly plan offerings are detailed here and range from free to $199 a month for 50,000 generative credits to create video and audio content at scale.