Skip to content
Mango9

Model help

Choose a model for your goal, check its required inputs, and understand the controls before generating.

A model is the AI engine that makes your image, video or sound. Choose it for the job and the source material you have. The suggestions below are starting points, not rankings or guarantees.

This selection changes the documentation only. Open Images in your project to generate. Availability and the final price are shown in Create.

For the guided script-and-speech workflow, including HeyGen Avatar IV and Kling Avatar choices, read Make a talking-avatar video. It is separate from the general video composer.

Black Forest Labs · Images · Create or edit an image

FLUX.2 Pro

Product photography and clean campaign visuals

A good starting point for

Start with one product reference when the packaging must remain recognizable. Describe the surface, light and camera angle, then inspect the label and shape at full size.

What you need

0–8 reference images

JPEG, PNG, WEBP · up to 10 MB each

Use FLUX.2 Pro in Create

  1. Open your project, then Images → Generate. Choose FLUX.2 Pro in the model dropdown.
  2. Prepare the inputs. 0–8 reference images. Wait for each upload to finish and check its thumbnail or audio preview. A completed upload is a source file, not a generated result.
  3. Write the prompt. Use the example below as a starting point and replace its subject with your own. For edits, say what to change and what to preserve.
  4. Check the controls and price. Use the supported choices in the form. Changing mode, duration or resolution can change other available fields and the quote. Review generation and price, then confirm once.
  5. Review the completed result. Find it in the project library or history. Enlarge it and inspect faces, product details and lettering. Regenerate as a new version to compare, or bring the approved result into Editor.

Try this example

Use the reference bottle unchanged. Place it on warm limestone beside a folded cream towel, soft morning window light, subtle shadow, uncluttered background.

Prompt length in this mode: 3–5,000 characters.

Images workspace with the model selector, input controls and output preview
The current Images workspace, captured November 15, 2026. The screenshot illustrates where the controls live; select FLUX.2 Pro to see this model’s own fields. Open the image to enlarge it.

Controls explained

These are the controls for the selected mode’s default settings. The live form narrows them as you change related choices; not every combination below is valid.

Resolution

Higher resolution creates a larger output. Check the price before generating.

1K · 2K
Aspect ratio

Choose the shape of the result. 9:16 is portrait; 16:9 is landscape.

16:9 · Landscape · 9:16 · Portrait · 1:1 · Square · 4:3 · 3:4 · 3:2 · 2:3

If the result is not what you expected

  • Generate is unavailable: check required files, prompt length, project access and any message under the form. A file in the library must still be selected in the input slot.
  • The model changed the subject: use a clearer reference and reduce competing instructions. Change one part of the prompt at a time.
  • The request is still running: check Project history and Notifications before starting another request. A second generation is a separate job.
  • The settings look different: check the exact model option and mode. Option 1 and Option 2 can expose different inputs.

Read the provider’s FLUX.2 Pro reference ↗
Reviewed 2026-11-06 against the provider reference and Create’s form. Providers may offer extra capabilities that this Create route does not expose. Examples and workflow advice are Mango9 guidance.

Read the complete images walkthrough →

The controls, in plain language

Prompt
The instructions for one result. For narration, the script contains the words to speak.
Reference image
A picture that guides appearance or style. It is not necessarily the first frame of a video.
Start / end frame
The composition where a clip begins or should finish. Keep both images compatible.
Aspect ratio
The shape: 9:16 is portrait, 16:9 is landscape and 1:1 is square.
Resolution
The pixel detail in the output. Higher resolution can change price and supported duration.
Duration
The length of one generated clip, not your entire edited project. Combine clips in Editor for a longer story.
Seed
A variation setting offered by some models. It does not guarantee an identical result across models or settings.
Option 1 / Option 2
Different routes to a model family. Compare inputs, settings and the live quote; they are not interchangeable forms.

First project: turn one image and a soundtrack into a reel →