Runflow
Black Forest Labs

FLUX.1 [schnell] API

FLUX.1 [schnell] — 12B-param speed-optimised model by Black Forest Labs. Sub-second generation at $0.003/MP. Apache 2.0 licensed.

flux-1-schnell12B paramsRectified Flow Transformer (distilled)
Updated March 2026·Early access
Price
$0.003/MP
Per megapixel · ~$0.003 for 1024×1024
Latency
Sub-second
1–4 inference steps, optimised for speed
Cold Start
5–10s
Not billed, handled by Runflow
Parameters
12B
Distilled Flow Transformer

FLUX.1 model family by Black Forest Labs

Different FLUX.1 variants for different speed, quality, and cost trade-offs — all on the same Runflow API.

FLUX.1 [schnell]

This model

Speed-optimised — sub-second generation at $0.003/MP, Apache 2.0

FLUX.1 [dev]

Standard

Open-weight, 12B params, $0.025/MP — production, LoRA, general purpose

FLUX.1 [pro]

Premium

$0.04/MP — commercial-grade, highest quality in the FLUX.1 family

Benchmarked for your use case

Every model on Runflow is scored per vertical — not generic leaderboard numbers. Scores come from domain-expert evaluation on real production prompts.

Use CaseAssessmentNotes
PortraitModerateFast drafts with decent quality — ideal for previews and iteration
FashionModerateGood for rapid concept exploration at high volume
ProductModerateQuick product shot drafts before committing to a premium model
CreativeStrongExcellent for brainstorming and rapid creative exploration

API Parameters

Full parameter reference for the flux-1-schnell model endpoint.

ParameterTypeDescription
promptstringText description of the image to generate.
image_sizeobject | stringOutput size. Presets: square_hd, square, portrait_4_3, portrait_16_9, landscape_4_3, landscape_16_9. Or custom {width, height}. Default: landscape_4_3.
num_inference_stepsintegerNumber of denoising steps. Range: 1–4. Default: 4. Lower = faster, higher = more detail.
guidance_scalefloatHow closely the model follows the prompt. Range: 1–20. Default: 3.5.
accelerationstringSpeed optimisation level. Options: "none", "regular", "high". Default: "regular". Higher = faster at minor quality cost.
num_imagesintegerNumber of images to generate per request. Range: 1–4. Default: 1.
output_formatstringOutput format. Options: "jpeg", "png". Default: "jpeg".
enable_safety_checkerbooleanEnable NSFW content filtering. Default: true.
seedintegerRandom seed for reproducible outputs.
sync_modebooleanReturn media as data URI without storing in request history. Default: false.

Technical Specifications

ArchitectureRectified Flow Transformer (distilled for speed)
Parameters12 billion
Inference Steps4 (default) · 1–4 range — 1–2 for fastest, 4 for balanced
Guidance Scale3.5 (default) · 1–20 range
Accelerationnone · regular · high — configurable speed boost
Max Batch Size4 images per request
Output FormatsJPEG · PNG
LicenseApache 2.0 — fully open, commercial use included

Why run FLUX.1 [schnell] on the Runflow API

Flux AI models are available on many platforms. Here's what makes the Runflow image generation API different.

Sub-second generation

FLUX.1 [schnell] is distilled for speed — 1 to 4 inference steps produce usable images in under a second, making it ideal for real-time previews and high-throughput pipelines.

$0.003 per megapixel

~8× cheaper than FLUX.1 [dev]. A 1024×1024 image costs $0.003. At this price point, you can generate thousands of drafts before committing to a premium model for final output.

Acceleration modes

The acceleration parameter (none/regular/high) lets you push speed even further when sub-second isn't fast enough — useful for live preview UIs and interactive design tools.

Apache 2.0 — fully open

No licensing restrictions. Use it commercially, fine-tune it, deploy it however you need. Weights and outputs are both unrestricted.

Sentinel quality scoring

Even at high speed, every output is scored by Runflow's Sentinel system before delivery, catching low-quality results automatically.

Explore other models

Same API, different strengths. Switch models with one parameter change.

Frequently asked questions

Resources

Start generating images with FLUX.1 [schnell]

Get API access in minutes. No GPU setup, no infrastructure to manage. Pair with Sentinel to control the quality of every AI-generated image you ship.