Runflow
Black Forest Labs

FLUX.1 [dev] API

FLUX.1 [dev] is a 12 billion parameter text-to-image model by Black Forest Labs built on a rectified flow transformer architecture. Known for strong prompt adherence, fine detail, and photorealism — it's one of the best open-weight Flux AI image generators available. Run it on Runflow with a single API call and pair with Sentinel to control quality on every output.

flux-1-dev12B paramsRectified Flow Transformer
Updated March 2026·Early access
Quality Score
Coming soon
Benchmarks in progress
Latency
Coming soon
Depends on steps & resolution
Price
Coming soon
Fixed per-image pricing
Parameters
12B
Rectified Flow Transformer

Flux AI model family by Black Forest Labs

Different Flux variants for different needs — all accessible through the same Runflow image generation API.

FLUX.1 [dev]

This model

Open-weight, 12B params, highest detail among open models

FLUX.2 Klein

Speed

Sub-second generation with state-of-the-art quality

FLUX.1 Kontext

Editing

In-context image generation and editing

FLUX.1 [dev] image generation examples

Real images generated with the FLUX.1 [dev] text-to-image model via the Runflow API — one API call per image, no post-processing.

AI-generated professional headshot of a woman with neutral grey background and soft studio lighting — created with FLUX.1 [dev] via Runflow API
Portrait

"Professional headshot of a woman, neutral grey background, soft studio lighting, sharp focus, corporate portrait"

AI-generated fashion editorial photo of a model wearing an oversized black blazer in a minimalist white studio — created with FLUX.1 [dev] via Runflow API
Fashion

"High-end fashion editorial, model wearing oversized blazer, minimalist white studio, dramatic side lighting"

AI-generated product photography of white sneakers on a reflective surface with clean commercial lighting — created with FLUX.1 [dev] via Runflow API
Product

"Product photography, white sneakers on reflective surface, clean background, commercial lighting"

AI-generated surreal landscape with floating islands at golden hour with cinematic composition — created with FLUX.1 [dev] via Runflow API
Creative

"Surreal landscape with floating islands, golden hour light, cinematic composition, 8k detail"

Benchmarked for your use case

Coming soon

Every model on Runflow is scored per vertical — not generic leaderboard numbers. Scores come from domain-expert evaluation on real production prompts. Benchmark results for FLUX.1 [dev] are currently in progress.

Use CaseQuality ScoreStatus
Portraits Benchmarking
Fashion Benchmarking
Product Benchmarking
Creative Benchmarking

API Parameters

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

ParameterTypeDescription
promptstringText description of the image to generate.
widthintegerImage width in pixels. Default: 1024.
heightintegerImage height in pixels. Default: 1024.
num_inference_stepsintegerNumber of denoising steps. Range: 10–50. Default: 50. Higher = better quality, slower.
guidance_scalefloatHow closely the model follows the prompt. Range: 1–20. Default: 3.5.
num_imagesintegerNumber of images to generate. Range: 1–4. Default: 1.
seedintegerRandom seed for reproducible results. Same seed + prompt = identical output.
output_formatstringOutput image format. Options: "jpeg", "png". Default: "jpeg".

Technical Specifications

ArchitectureRectified Flow Transformer
Parameters12 billion
Default Resolution1024 × 1024
Inference Steps50 (default) · 10–50 range
Guidance Scale3.5 (default) · 1–20 range
Max Batch Size4 images per request
Output FormatJPEG (default), PNG
LicenseFLUX.1 [dev] Non-Commercial License — outputs usable commercially when generated via hosted API

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

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

Benchmarked per niche

We don't give you a generic quality number. Every model is scored for portraits, fashion, product, and creative — so you pick the right model for your use case, not the hyped one.

Fixed per-image pricing

No GPU-second billing. No cold-start surcharges. Simple fixed pricing per image. Predictable costs at any scale. Pricing details coming soon.

Sentinel: control every output

AI generation isn't deterministic — some outputs miss the mark. Sentinel evaluates every image before delivery, scoring against your criteria. Bad outputs get caught and retried automatically. Your customers only see quality.

Learn more

Built for reliability

Multi-region GPU infrastructure with automatic failover. Your pipeline doesn't go down because one datacenter has a bad day.

One API for every model

Switch from FLUX.1 to Seedream to SDXL by changing one parameter. No vendor lock-in, no different SDKs, no rewriting integrations.

Production-tested

Built by a team that ran 100,000+ inference jobs in production before launching the platform. Not a wrapper — real infrastructure.

Learn more

Explore other models

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

FLUX.1 Kontext

In-context image generation and editing

Editing

FLUX.2 Klein

Sub-second generation, state-of-the-art quality

Speed

Seedream 4.5

Strong performer for product photography

Product

Reve

High-quality creative image generation

Creative

Qwen Image Max

Versatile multimodal image model

Multimodal

Kling 3.0

Video generation from text and image

Video

Frequently asked questions

Resources

Start generating images with FLUX.1 [dev]

Get Flux 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.