Black Forest Labs
Sub-Second Inferenceby Black Forest Labs

FLUX.1 Schnell

Black Forest Labs' speed-optimized FLUX variant — sub-second image generation at $0.003 per megapixel. Distilled 12B parameter flow transformer that prioritizes speed without giving up FLUX's photorealism character.

Model Specs

Released
Aug 2024
Aspect ratios
5 options
Output formats
PNG, JPEG, WEBP
Modalities
textvision
What it can produce

Image generation capabilities

Aspect ratios, output formats, and style options supported by this model.

Aspect ratios
5
Square (1024×1024)Landscape (1024×768)Portrait (768×1024)Wide Landscape (1536×1024)Tall Portrait (1024×1536)
Output formats
3
PNGJPEGWEBP
Quality tiers
1
Standard

About this model

FLUX.1 Schnell is Black Forest Labs' speed-optimized variant of the FLUX.1 family — 'schnell' is German for 'fast', and the model lives up to the name. Through flow-based distillation from the original FLUX.1 dev/pro models, Schnell achieves sub-second inference times (0.366 seconds in fal.ai's example) at 1-4 inference steps, compared to the typical 20-50 steps of standard diffusion models. The architecture is a 12-billion-parameter flow transformer — the same scale as FLUX.1 dev, just optimized for speed.

Pricing on fal.ai is $0.003 per megapixel — making Schnell among the cheapest image generation options on Renas AI. A standard 1MP (1024×1024) generation costs $0.003 raw, which is roughly 10x cheaper than FLUX.2 Pro's $0.03/MP base rate. Output is configurable as PNG (default) or JPEG, and the model supports both text-to-image and image-to-image generation. Custom width/height configurations are supported, with the example at 1024×768.

On Renas AI, FLUX.1 Schnell is available in the AI Image tool. Reach for it when (a) you're generating large volumes of images and per-image cost compounds, (b) speed matters — interactive generation, real-time preview, rapid iteration, (c) you're A/B testing prompts before committing to a higher-quality regeneration on FLUX.2 Pro, or (d) you want FLUX's photorealistic output character at the budget tier. For production hero assets and brand-defining work, escalate to FLUX.2 Pro or Nano Banana Pro.

Key Strengths

Sub-second inference

Fal.ai's example completed in 0.366 seconds. For interactive workflows where waiting time breaks the flow, Schnell's speed transforms the experience — generate, see, iterate without breaking concentration.

Cheapest FLUX-family pricing

$0.003 per megapixel — 10x cheaper than FLUX.2 Pro. A standard 1MP generation costs $0.003 raw. Makes high-volume FLUX workflows economical.

FLUX photorealism at budget tier

Schnell is distilled from the original FLUX.1, not a separate smaller architecture. The photorealistic output character that defines the FLUX family carries through to Schnell, just at slightly lower fidelity than the Pro variants.

1-4 step inference

Distillation enables high-quality output in just 1-4 inference steps (default 4) vs the typical 20-50 of standard diffusion models. Default 4 steps balances speed and quality; lower steps for maximum speed.

Text-to-image + image-to-image

Both generation modes supported. T2V for new generations from prompts, I2I for variations on existing images at minimal cost.

12B parameter architecture

Schnell isn't a small model — it's a 12-billion-parameter flow transformer. The speed comes from the distillation process, not from a smaller backbone, which is why output quality holds up.

How it compares

FLUX.1 Schnell competes with budget photorealism models from other providers. Pick by output character preference and ecosystem.

vs. ModelVerdictOutcome

Pros

  • Sub-second inference (0.366s example)
  • Cheapest FLUX-family pricing ($0.003/MP)
  • FLUX photorealism character at budget tier
  • 12B parameter architecture — no quality compromise from smaller backbone
  • Both text-to-image and image-to-image supported
  • Same prompt-engineering style as FLUX.2 Pro — easy promotion path

Things to consider

  • Lower fidelity than FLUX.2 Pro on complex prompts
  • No JSON structured prompts (Pro feature)
  • No HEX color brand control (Pro feature)
  • No image editing or style transfer (Pro feature)
  • Best as drafting tool, not final-asset generator
  • Per-megapixel pricing scales cost for high-resolution output

Best use cases

Rapid prompt iteration

Generate 10-20 variations of a concept in seconds at $0.003 each, pick the strongest, then either ship Schnell output or regenerate the winner on FLUX.2 Pro for production polish.

High-volume content generation

Bulk product imagery, social media post variations, blog thumbnails. The $0.003/MP rate makes scale economical without leaving the FLUX ecosystem.

Real-time / interactive applications

Sub-second inference enables UX patterns where speed matters — live previews as users type, interactive editing where each step generates visual feedback.

Concept exploration

Mood boards, brainstorming visuals, design ideation. Schnell's price + speed combo lets you explore broadly before committing to a final direction.

Provider comparison testing

Generate the same prompt on Schnell, GPT Image Mini, Imagen 4 Fast, and Nano Banana — see which budget photorealism style fits your brand. Schnell's FLUX character is the differentiator.

Educational and explanatory visuals

Course material illustrations, tutorial supporting images. Schnell's output is sufficient for visuals that support text rather than carrying the message alone.

How to use it on Renas AI

  1. 1

    Step 1

    Open the AI Image tool

    Navigate to AI Image in the Renas dashboard. Pick FLUX.1 Schnell from the model selector — it's marked as the speed-optimized FLUX variant. The tool shows live credit cost based on output dimensions.

  2. 2

    Step 2

    Pick output dimensions

    Schnell supports preset sizes or custom dimensions. Standard 1024×1024 (1MP) is a solid default for most workflows. For larger outputs, the per-megapixel pricing scales linearly.

  3. 3

    Step 3

    Write a clear FLUX-style prompt

    FLUX prompt style works well — describe subject, lighting, camera angle, mood. Schnell uses the same prompt patterns as other FLUX variants, so prompt engineering carries over from Pro/dev workflows.

  4. 4

    Step 4

    Iterate fast — escalate when ready

    Schnell's speed and price reward rapid iteration. Test 5-10 prompt variations cheaply, find the winning concept, then either ship Schnell output or regenerate on FLUX.2 Pro at $0.03/MP for production-quality finals.

Pricing

Pricing on Renas AI

Pay-as-you-go credits, no API keys, no rate limits.

10credits per image
Included in every paid plan
No separate API key or setup
Predictable per-word credit cost
Commercial use rights for all output

Frequently asked questions

Sub-second AI image generation

Use FLUX.1 Schnell with your Renas AI subscription credits — no API key, no setup, no per-seat fees.

Try FLUX.1 Schnell
FLUX.1 Schnell by Black Forest Labs — Sub-Second AI Image | Renas AI | Renas AI