FLUX Models Compared: FLUX.1 [dev] vs Schnell vs FLUX.2 Pro and More
Which FLUX model to use: FLUX.1 [dev], Schnell, FLUX.2 Pro, 1.1 Pro Ultra and Kontext Max compared by task, inputs and cost, plus a simple workflow.
FLUX is a family of AI image models from Black Forest Labs and one of the most popular choices for photorealistic image generation. The name covers several quite different tools: the original FLUX.1 models (often searched as flux 1 or flux.1 ai), the FLUX 1.1 Pro line, the newer FLUX.2 generation and the Kontext editors. This guide compares the FLUX models available on WebTerra AI — FLUX.1 [dev], FLUX Schnell, FLUX.2 Pro, FLUX 1.1 Pro Ultra and FLUX Kontext Max — and explains which one to pick for which job.
All of them run online in the same WebTerra AI image generator, so you can test one prompt on several models and compare the results side by side, without installing anything or renting a GPU.
The FLUX family at a glance
Black Forest Labs released FLUX.1 in 2024 in three variants: [pro], [dev] and [schnell]. The [dev] and [schnell] versions were published with open weights, which is a big part of why FLUX spread so quickly through community tools. Since then the family has grown: FLUX 1.1 Pro and its high-resolution Ultra version, the Kontext line for editing photos with text instructions, and FLUX.2 as the newer generation focused on photorealism. On WebTerra AI each of them is a separate model card, and the dashboard shows the price before every run.
| Model | Type | Input | Best for | Relative cost |
|---|---|---|---|---|
| FLUX Schnell | Text-to-image | Text, up to 1 reference | Drafts, tests, bulk generation | Lowest |
| FLUX.1 [dev] | Text-to-image, image-to-image | Text, 1 reference with adjustable strength | Everyday volume, photo stylization | Low |
| FLUX.2 Pro | Generation and editing | Text, up to 4 references | Photorealism, portraits, product shots | Low |
| FLUX 1.1 Pro Ultra | Text-to-image, image-to-image | Text, 1 reference, raw mode | 4 MP images for print and banners | Medium |
| FLUX Kontext Max | Instruction-based editor | Photo + text, up to 4 images | Complex multi-step photo edits | Higher |
“Relative cost” compares the five models with each other at default settings. Actual prices depend on your plan and the parameters you choose, so see the price on the model card before you generate.
The FLUX models one by one
FLUX Schnell: fast and cheapest
FLUX Schnell, officially FLUX.1 [schnell] and also searched as flux 1 schnell, is the speed-optimized model of the original FLUX.1 family; “schnell” is German for “fast”. It turns a text prompt into an image and accepts one optional reference. Its job is exploration: test an idea, try several compositions, generate dozens of variants, and only then move the winning prompt to a bigger model. The trade-off is detail: fine textures, tricky hands and lettering are better handled by the larger FLUX models. Read the full FLUX Schnell review.
FLUX.1 [dev]: the predictable workhorse
FLUX.1 [dev], searched as flux 1 dev, flux.1-dev or black-forest-labs/FLUX.1-dev, is a time-tested model that follows prompts reliably and renders rich detail at a low price. Besides text-to-image it supports image-to-image with adjustable strength: upload one photo and decide whether the model should only nudge the style or rework the picture. That makes it a sensible choice for a steady stream of social media visuals, blog illustrations, mood boards and stylized versions of your own photos. See the FLUX.1 [dev] review.
FLUX.2 Pro: photorealism and editing with references
FLUX.2 Pro, also written as Flux 2 Pro or FLUX 2.0 Pro, belongs to the newer generation and specializes in photorealism, portraits and artistic styles. It works in two modes: generation from text, and prompt-based editing with up to four reference images — for example, moving a product into a new scene while keeping its logo and colors. If you need realistic people, products, food or interiors, start here. Details in the FLUX.2 Pro review.
FLUX 1.1 Pro Ultra: 4 megapixels without upscaling
FLUX 1.1 Pro Ultra generates 4-megapixel images straight away, so the detail is created during generation rather than invented by an upscaler afterwards. You set the size through the aspect ratio, can add one reference image and switch on raw mode for a less polished, more natural look. It is the model for print, large banners and catalogs. If 4 MP is not needed, the regular FLUX 1.1 Pro is the proven classic for everyday work. Full breakdown in the FLUX 1.1 Pro Ultra review.
FLUX Kontext Max: edit a photo with words
FLUX Kontext Max, officially FLUX.1 Kontext [max], is not a from-scratch generator but an editor. You upload a photo (up to four images per request) and describe the change: remove an object, replace the background, retouch a portrait, add a caption. Max understands multi-step instructions and keeps the original composition — angle, proportions and object placement. For one simple fix, the younger FLUX Kontext Pro is the cheaper option. More in the FLUX Kontext Max review.
Which FLUX model to choose for your task
- Testing ideas, mood boards, dozens of variants — FLUX Schnell.
- Regular social posts, blog illustrations, budget volume — FLUX.1 [dev].
- Turning your own photo into a watercolor or illustration — FLUX.1 [dev] with moderate image-to-image strength.
- Realistic portraits, product shots, food and interiors — FLUX.2 Pro.
- Putting a product into a new scene using several references — FLUX.2 Pro, up to 4 references.
- Posters, catalogs, large banners and print — FLUX 1.1 Pro Ultra.
- Removing objects, swapping backgrounds, several edits at once — FLUX Kontext Max.
- One quick fix on an existing photo — FLUX Kontext Pro.
FLUX.1 [dev] vs FLUX Schnell
These two are the most frequently compared FLUX models, because both come from the original FLUX.1 family and both are budget-friendly. The differences are practical:
- Speed and price. Schnell is the fastest and cheapest FLUX model in the catalog; [dev] costs more per image.
- Detail. [dev] renders richer textures and small scene elements; Schnell trades some of that for speed.
- Control over your photo. [dev] offers image-to-image with a strength setting; Schnell accepts a reference but is mainly a text-to-image tool.
- Licence of the open weights. [schnell] is released under the permissive Apache 2.0 licence, while [dev] uses a non-commercial licence for the model itself.
A practical rule: explore in Schnell, produce usable volume in [dev], and move the images that need top realism to FLUX.2 Pro.
FLUX.1 vs FLUX.2: what changed
FLUX.1 is the original generation: [dev] and [schnell] are the affordable workhorses, and FLUX 1.1 Pro is the established classic. FLUX.2 Pro is positioned as an all-rounder of the newer generation, with a stronger focus on photorealism and on editing with several references at once. In practice you don't have to commit to one generation. The older models remain the economical choice for drafts and volume, while FLUX.2 Pro is where you go when people and products need to look as real as possible.
A cost-efficient FLUX workflow
- Explore in FLUX Schnell. Write three or four different takes on the idea and generate each a few times.
- Lock the prompt. Describe subject, setting, light, style and camera, and change one detail at a time.
- Produce. Generate volume in FLUX.1 [dev] or the hero shot in FLUX.2 Pro.
- Go big. If the image is headed for print or a large banner, render it in FLUX 1.1 Pro Ultra.
- Polish. Use FLUX Kontext Max to remove a stray object, swap the background or add a caption without regenerating the whole frame.
This keeps the more expensive generations for ideas you have already tested. Write prompts in English: the FLUX reviews consistently note that the models interpret English most precisely. For a step-by-step method, see how to write prompts for AI images.
When FLUX is not the best choice
FLUX models are strongest in photorealism and visual detail, but no family is best at everything. When the image must contain legible text — a poster headline, a label, a logo — look at Nano Banana 2 or Ideogram V3. For vector-style logos and icons, Recraft V3 is the specialist. And for dense scenes with many characters and props, compare FLUX with Hunyuan Image 3.
How FLUX model names are written
The same models appear under many spellings, which can be confusing when you search:
- FLUX.1 [dev] — flux 1 dev, flux.1-dev, flux1-dev, black-forest-labs/FLUX.1-dev (the name of its Hugging Face repository).
- FLUX.1 [schnell] — FLUX Schnell, flux 1 schnell, flux-1-schnell, schnell flux.
- FLUX.2 Pro — Flux 2 Pro, FLUX 2.0 Pro.
- FLUX 1.1 Pro Ultra — FLUX1.1 [pro] Ultra, flux 1.1 ultra.
- FLUX Kontext Max — FLUX.1 Kontext [max].
In short: “FLUX.1” is the first generation, “1.1” its updated Pro version, and “FLUX.2” the newer generation.
FAQ
Which FLUX model is the best?
It depends on the job. For realistic portraits and products, FLUX.2 Pro; for print-size images, FLUX 1.1 Pro Ultra; for editing an existing photo, FLUX Kontext Max; for cheap drafts and volume, FLUX Schnell or FLUX.1 [dev].
Is FLUX free?
The FLUX.1 [dev] and [schnell] weights are openly available from Black Forest Labs, so you can run them yourself on a capable GPU — [schnell] under Apache 2.0, [dev] under a non-commercial licence. The Pro, Ultra and Kontext Max models are offered through paid APIs. On WebTerra AI you need no hardware: you pay per image, and the price is shown on the model card before each run.
What is FLUX.1 AI?
FLUX.1 is the first generation of FLUX image models from Black Forest Labs, released in 2024. When people search for “flux.1 ai” or a “flux 1 image generator”, they usually mean FLUX.1 [dev] or FLUX.1 [schnell].
Can I use images from FLUX models commercially?
The rules depend on the model licence and on the terms of the service you generate with. Before a commercial project, read the WebTerra AI terms of service and check the result for details such as hands, logos and small text.
Do I need a powerful computer?
No. On WebTerra AI the models run in the cloud: open the dashboard in a browser on desktop or mobile, choose a model and type a prompt.
All FLUX models, along with other image generators, are collected on the AI image generation page. Pick the one that fits your task, compare a prompt across two or three models and create your first image in WebTerra AI.
Try AI models yourself
Models for video, images and audio in one place — a free trial balance when you sign up
Start for free

