Black Forest Labs' second generation image family generates and edits pictures from text and multiple references in one model. Four text to image variants from 1 credit, three edit variants, five aspect ratios, no watermark. Available on every Aura AI plan.
FLUX 2, written FLUX.2 by its maker, is the second generation of the FLUX image models from Black Forest Labs. It was announced on November 25, 2025 as a family rather than a single checkpoint: FLUX.2 [pro] for top quality through the API, FLUX.2 [flex] with exposed step and guidance controls, FLUX.2 [dev] as a 32 billion parameter open-weight model, and FLUX.2 [klein], a size-distilled open model that followed in January 2026. A FLUX.2 Max tier sits above Pro in the vendor lineup.
The main change compared to the original FLUX 1 Dev is that generation and editing now live in the same model. FLUX 2 accepts a text prompt plus up to 10 reference images at once upstream, edits at up to 4 megapixels, and Black Forest Labs specifically calls out reliable fine text, so typography, infographics and UI mockups no longer need a separate pass. The architecture pairs a Mistral 3 24B vision language model with a rectified flow transformer.
On Aura AI, FLUX 2 runs in text to image mode as FLUX 2, FLUX 2 Pro, FLUX 2 Max and FLUX 2 Turbo, and in the image editor as FLUX 2 Edit, FLUX 2 Pro Edit and FLUX 2 Max Edit. Every variant costs a flat 1 to 3 credits per image, outputs in five aspect ratios and is delivered without watermark. It sits next to FLUX Kontext, Nano Banana and GPT Image 2 in the same selector.
Six capabilities that separate FLUX 2 from the first generation
Upstream, FLUX 2 reads as many as 10 reference images in a single request. Lock a face, a product and a logo at the same time and the model keeps all of them consistent in the output.
The first FLUX generation split text to image (FLUX 1) from editing (Kontext). FLUX 2 does both from the same weights, so a picture you generate can be edited with the same look.
Black Forest Labs quotes editing at up to 4 megapixels, enough for print proofs, hero banners and product shots that need to survive cropping without an upscaler.
Typography, infographics, memes and UI mockups with small text are listed by the maker as reliable use cases. Label your text in quotes in the prompt and FLUX 2 renders it readable at small sizes.
Photorealism, image reference fidelity and prompt understanding are the four areas Black Forest Labs lists as improved over FLUX 1, together with typography. Product photography and lifestyle imagery are where the difference shows most.
On Aura AI, FLUX 2 Turbo costs a single credit per image, the same as FLUX 1 Dev. It is the fastest way to draft compositions before spending 2 or 3 credits on a Pro or Max render.
Four steps from prompt to a finished image
Flat cost per image, identical for every aspect ratio
| Variant | Mode | Credits per image | Best for |
|---|---|---|---|
| FLUX 2 Turbo | Text to image | 1 credit | Fast drafts, thumbnails, batch ideation |
| FLUX 2 | Text to image | 2 credits | Everyday generation, social visuals |
| FLUX 2 Pro | Text to image | 2 credits | Photorealistic renders, product shots |
| FLUX 2 Max | Text to image | 3 credits | Highest fidelity, fine text, print |
| FLUX 2 Edit | Image edit | 2 credits | Quick local changes on an existing image |
| FLUX 2 Pro Edit | Image edit | 2 credits | Reference-driven edits, character consistency |
| FLUX 2 Max Edit | Image edit | 3 credits | Demanding edits with text or many references |
How the new Black Forest Labs family compares to its siblings on Aura AI
| Feature | FLUX 2 | FLUX 1 Dev | Nano Banana | GPT Image 2 |
|---|---|---|---|---|
| Maker | Black Forest Labs | Black Forest Labs | OpenAI | |
| Text to image credits | 1 to 3 | 1 | 3 | 4 |
| Image edit credits | 2 to 3 | Via FLUX Kontext | 2 to 3 | 4 |
| Cheapest variant | FLUX 2 Turbo, 1 credit | FLUX 1 Dev, 1 credit | Nano Banana Edit, 2 credits | GPT Image 2, 4 credits |
| Multi-reference editing | ✓ Up to 10 images (upstream) | No | ✓ Up to 14 images (Pro, upstream) | ✓ Yes |
| Open weights | ✓ FLUX.2 [dev], [klein] | ✓ Yes | No | No |
| Aspect ratios on Aura AI | 5 | 5 | 5 | 5 |
| Plan | All plans | All plans | All plans | All plans |
Where multi-reference editing and fine text pay off
Upload one photo of your product in the editor and place it in a new scene. FLUX 2 Pro Edit keeps the shape, label and colour of the original while changing the background, surface and lighting.
FLUX 2 Max renders small text legibly, which makes it the pick for posters, menu boards and social infographics where the words are part of the picture.
Generate a character once, then feed the result back as a reference for every new scene. The same face, outfit and proportions carry through a comic strip, a storyboard or a brand mascot series.
Draft app screens and ad layouts with FLUX 2 Turbo at 1 credit, then re-render the best one with FLUX 2 Max for the final asset.
Common questions about FLUX 2 on Aura AI
Text to image and multi-reference editing, from 1 credit per image, no watermark. Available on every plan, worldwide.