How much does Flux Schnell cost via Railwail?

Per-call: €0.03. No monthly minimum, no subscription. Start with €5 free credits.

What is the context window of Flux Schnell?

Flux Schnell supports a unknown context window — enough for typical AI workloads.

How fast is Flux Schnell?

Average response latency: 2.0s (p50 across recent Railwail traffic). See live p50/p95 metrics on /rankings.

Is Flux Schnell better than Flux 1.1 Pro Ultra?

It depends on your use case. Flux Schnell (Black Forest Labs) and Flux 1.1 Pro Ultra (Black Forest Labs) are both strong choices in image generation. Compare them side-by-side at /compare/flux-schnell-vs-flux-pro-ultra.

Flux Schnell

Name: Flux Schnell
Brand: Replicate
SKU: flux-schnell
Price: 0.03 EUR
Availability: InStock

Black Forest Labs

Image Generation

The fastest Flux model. Generate images in under 2 seconds. Great for prototyping.

Generate with Flux Schnell

Describe what you want and pick a size — the image renders inline.

Size

Result appears here

TL;DR·Last updated March 4, 2026

Flux Schnell is image generation AI model from Black Forest Labs, priced at €0.000 per 1M input tokens with a unknown context window.

Try Flux Schnell

Prompt

Aspect Ratio

Quality

Examples

See what Flux Schnell can generate

Sample output

Quick Concept

Prompt: "A friendly robot barista serving coffee in a cozy cafe, warm lighting, cartoon illustration style with soft pastel colors"

Sample output

Icon Design

Prompt: "A minimalist flat design app icon of a rocket ship launching from a phone screen, gradient purple to blue background, clean vector style"

Sample output

Food Photography

Prompt: "A rustic wooden board with assorted artisan cheeses, figs, honey drizzle, and crusty bread, overhead shot with dramatic side lighting on a dark slate surface"

Pricing

Price per Generation

Per generation€0.03

API Integration

Use our OpenAI-compatible API to integrate Flux Schnell into your application.

Install

npm install railwail

JavaScript / TypeScript

import railwail from "railwail";

const rw = railwail("YOUR_API_KEY");

const images = await rw.run("flux-schnell", "A beautiful sunset over Tokyo");
console.log(images[0].url);

// Or use the image() method for full control
const res = await rw.image("flux-schnell", "A cat in space", {
  size: "1024x1024",
  n: 1,
});
console.log(res.data[0].url);

Specifications

Price

€0.03

Avg. latency

2.0s

Est. duration

Developer

Black Forest Labs

Deep dive — Black Forest Labs's Flux Schnell

About Black Forest Labs

Founded 2024 · Freiburg, Germany

Black Forest Labs was founded in August 2024 in Freiburg, Germany by Robin Rombach, Patrick Esser, Andreas Blattmann and Dominik Lorenz, the original Latent Diffusion and Stable Diffusion team. After raising a $31M seed round led by Andreessen Horowitz, the lab launched the FLUX.1 family. The 'schnell' tier (German for 'fast') is positioned as the high-volume, low-latency commercial offering, available both as Apache 2.0 weights and as a commercial API endpoint on Black Forest Labs, Replicate, fal.ai and Together AI.

Visit Black Forest Labs →

Architecture

Rectified-flow Diffusion Transformer (DiT) distilled to 1-4 steps

FLUX.1 [schnell] is the few-step distilled variant of the FLUX.1 family, optimised for sub-second inference. Architecturally it is the same 12B rectified-flow DiT as FLUX.1 [dev], using T5-XXL plus CLIP-L text encoders and a 16x latent VAE. Schnell is produced by adversarial latent-space distillation in the spirit of Adversarial Diffusion Distillation (Sauer et al. 2023), reducing the required sampling steps from 28-50 down to 1-4 with a small but measurable quality drop. The commercial schnell endpoint on Black Forest Labs and partner platforms exposes 1, 2 and 4 step modes, with 4 steps recommended for production. The model is also available open-weights under Apache 2.0 (see 'flux-1-schnell-apache'). The commercial tier additionally includes input/output safety filters and SLA support.

Parameters: ~12B parameters
Context: 512 tokens

What it can do

1-4 step sampling — sub-second per 1MP image
Strong photorealism and prompt adherence
Up to 2 MP output resolution
Decent in-image text rendering
Available as Apache 2.0 weights for self-hosting
Commercial API tier with SLA and safety filters
Compatible with FLUX Tools and LoRA ecosystem
Best for: real-time creative apps, batch generation, chat-based image generation, mobile UIs.

Training & License

Same base pretraining as FLUX.1 [dev]/[pro] on a large licensed image-text corpus, followed by adversarial latent-space distillation to enable few-step sampling.

License: Dual: Apache 2.0 for the open weights (commercial use permitted); proprietary commercial license for the hosted API tier.

Known limitations