Blog/Model Guide
Model Guide

How to Use FLUX.2: The Practical Guide (2026)

By the Shotari Team ·

FLUX.2 has a reputation to live up to: it comes from the Black Forest Labs team whose research lineage produced Stable Diffusion, and its calling card is images that read as photographs rather than renders. This guide covers what it is genuinely best at, how to run it in a browser with no setup, and the prompt patterns that get the most out of it — including where a different model is honestly the better pick.

July 18, 2026 · 6 min read

What Is FLUX.2, Actually?

FLUX.2 is Black Forest Labs' flagship text-to-image model, released in November 2025 as the successor to the FLUX.1 family that dominated open-model leaderboards through 2024–2025. Three things define it in practice: photorealism — skin, fabric, and light render with a naturalness that made "the FLUX look" a shorthand; prompt adherence — long, specific prompts come back respected instead of averaged away; and native image input — it accepts reference images alongside text, so it edits and restyles as well as it generates.

The catch, historically, was access: running FLUX yourself means GPUs and setup. On Shotari's FLUX.2 page it runs in the browser at 6 credits per image — no install, same account and credit balance as every other model.

How to Run FLUX.2 in Your Browser

1

Open the FLUX.2 page

Go to shotari.com/flux-2 — the composer preselects FLUX.2. (It's also in the model picker on the workspace and generator pages.)

2

Write the full brief

FLUX.2 rewards specificity. Describe subject, lighting, lens feel, and mood in one prompt — see the patterns below.

3

Attach references if editing

Up to 3 images to lock identity, style, or composition. Skip this for pure text-to-image.

4

Pick resolution and generate

Draft at 1K (6 credits), ship at 2K or 4K. Results land in your history for side-by-side comparison.

Prompt Patterns That Play to Its Strengths

FLUX.2 responds unusually well to photography language. Instead of style keywords bolted together, describe the shot the way a photographer would brief it: light source and direction, lens feel, texture, wardrobe, mood. Copy-paste starters:

Editorial portrait

Portrait of a woman in her 60s with silver hair, window light from the left, shallow depth of field, 85mm lens look, natural skin texture, muted earth-tone wardrobe, quiet confidence

Product hero shot

A matte-black wireless headphone on a concrete pedestal, dramatic side lighting, soft gradient background from charcoal to black, subtle reflection, premium tech advertising style

Food photography

Overhead shot of a rustic sourdough loaf on a linen cloth, scattered flour, warm morning light through a window, steam rising, shallow focus on the crust

Poster with text

Minimalist concert poster with the headline 'MIDNIGHT SESSIONS' in bold condensed type, a saxophone silhouette in deep blue, grainy texture, Swiss design layout

Two habits pay off: name the light ("window light from the left", "golden hour rim light") — it is the single biggest realism lever; and for text in images, put the exact words in quotes and keep them short. For general prompt structure, our prompt-writing guide applies fully here.

FLUX.2 vs the Other Models on Shotari

Pick…When…Credits
FLUX.2the image must read as a photograph — portraits, products, food, editorial6
Nano Banana Prothe same subject must stay consistent across a series10
GPT Image 2you're iterating fast on complex prompts and want volume5
Seedream 5volume economics matter most — social pipelines, catalog work3

Deeper head-to-heads live in the model comparison hub — start with Nano Banana Pro vs GPT Image 2if you're choosing a daily driver.

FAQ

What is FLUX.2 and who makes it?

FLUX.2 is the flagship image generation model from Black Forest Labs, released in November 2025. The team's lineage traces back to the original Stable Diffusion research, and FLUX.2 is best known for photorealistic rendering and close prompt adherence.

How much does FLUX.2 cost to use?

On Shotari it's 6 credits per 1K image (2K costs 1.5×, 4K costs 2×). That sits mid-ladder: cheaper than Nano Banana Pro at 10 credits, slightly above GPT Image 2 at 5. Credit packs start at $29.90 and subscriptions at $20/mo.

Does FLUX.2 support reference images?

Yes — text+image input is native to FLUX.2. On Shotari you can attach up to 3 reference images to guide identity, composition, or style alongside your prompt.

Is FLUX.2 better than Nano Banana Pro?

Different strengths: FLUX.2 leads on single-image photorealism at a lower price; Nano Banana Pro leads on keeping a subject consistent across a whole series. Draft striking one-offs on FLUX.2; build character or brand series on Nano Banana Pro.

Can I use FLUX.2 images commercially?

On Shotari, images generated with paid credits are watermark-free with full commercial usage rights. Free-tier output carries a small watermark for personal use.