FLUX.2 [dev]

Name: FLUX.2 [dev]
Brand: Black Forest Labs
Price: 0.012 USD
Availability: InStock

AI Image Editing Model

Image

Black Forest Labs' open-weights image generation model with frontier performance, available for non-commercial local deployment

Try Model Benchmark examples

2048 x 2048

Max Resolution

Supported Modes

Text to Image Image Edit

Active

Details

Model ID

flux.2-dev

Creator

Black Forest Labs

Family

flux.2

Max Input Images

Ready to integrate?

Access flux.2-dev via our unified API.

Create Account

Available at 2 providers

Starting from

$0.012 /image via fal.ai · +1 more

Prices shown are in USD · Some prices estimated from per-megapixel or per-token pricing

See all providers

Provider Performance

Fastest generation through replicate at 5,827ms median latency with 84.6% success rate.

Aggregated from real API requests over the last 30 days.

Generation Time

replicate

p95: 11,907ms

fal

p95: 12,277ms

Success Rate

replicate

84.6%

11 / 13 requests

fal

100.0%

16 / 16 requests

Time to First Byte

replicate

5,293ms

p95: 11,199ms

fal

8,396ms

p95: 11,607ms

Provider Rankings

#	Provider	p50 Gen Time	p95 Gen Time	Success Rate	TTFB (p50)
1	replicate	5,827ms	11,907ms	84.6%	5,293ms
2	fal	8,089ms	12,277ms	100.0%	8,396ms

Data updated every 15 minutes. Based on all API requests through Lumenfall over the last 30 days.

Providers & Pricing (3)

FLUX.2 [dev] is available from 3 providers, with per-image pricing starting at $0.012 through fal.ai.

fal.ai

Text to Image

fal/flux.2-dev

Provider Model ID: fal-ai/flux-2

$0.012 /megapixel

View pricing details

fal.ai

Image Edit

fal/flux.2-dev-edit

Provider Model ID: fal-ai/flux-2/edit

Input

Megapixel

$0.012

Output

Megapixel

$0.012

Pricing Notes (4)

• Resolution is rounded up to the next megapixel, separately for each reference image and the generated image
• 1 megapixel = 1024x1024 pixels
• Each reference image is counted separately (minimum 1 MP each)
• Images exceeding 4 megapixels are resized to 4 megapixels

View official pricing • As of Dec 28, 2025

Replicate

Text to Image Image Edit

replicate/flux.2-dev

Provider Model ID: black-forest-labs/flux-2-dev

Input

Megapixel

$0.014

Output

Megapixel

$0.014

Pricing Notes (2)

• Resolution is rounded up to the next megapixel, separately for each reference image and the generated image
• 1 megapixel = 1024x1024 pixels

View official pricing • As of Dec 30, 2025

FLUX.2 [dev] API OpenAI-compatible

Integrate FLUX.2 [dev] through Lumenfall’s unified API to programmatically generate or edit images using an OpenAI-compatible interface. This allows developers to trigger high-quality text-to-image generations and image-to-image transformations with a single standardized request.

Base URL

https://api.lumenfall.ai/openai/v1

Model

flux.2-dev

Code Examples

Text to Image

/v1/images/generations

curl -X POST \

  https://api.lumenfall.ai/openai/v1/images/generations \

  -H "Authorization: Bearer $LUMENFALL_API_KEY" \

  -H "Content-Type: application/json" \

  -d '{

    "model": "flux.2-dev",

    "prompt": "",

    "size": "1024x1024"

}'

# Response:

# { "created": 1234567890, "data": [{ "url": "https://...", "revised_prompt": "..." }] }

import OpenAI from 'openai';

const client = new OpenAI({

  apiKey: 'YOUR_API_KEY',

  baseURL: 'https://api.lumenfall.ai/openai/v1'

});

const response = await client.images.generate({

  model: 'flux.2-dev',

  prompt: '',

  size: '1024x1024'

});

// { created: 1234567890, data: [{ url: "https://...", revised_prompt: "..." }] }

console.log(response.data[0].url);

from openai import OpenAI

client = OpenAI(

    api_key="YOUR_API_KEY",

    base_url="https://api.lumenfall.ai/openai/v1"

response = client.images.generate(

    model="flux.2-dev",

    prompt="",

    size="1024x1024"

# { created: 1234567890, data: [{ url: "https://...", revised_prompt: "..." }] }

print(response.data[0].url)

Image Edit

/v1/images/edits

curl -X POST \

  https://api.lumenfall.ai/openai/v1/images/edits \

  -H "Authorization: Bearer $LUMENFALL_API_KEY" \

  -F "model=flux.2-dev" \

  -F "[email protected]" \

  -F "prompt=Add a starry night sky to this image" \

  -F "size=1024x1024"

# Response:

# { "created": 1234567890, "data": [{ "url": "https://...", "revised_prompt": "..." }] }

import OpenAI from 'openai';

import fs from 'fs';

const client = new OpenAI({

  apiKey: 'YOUR_API_KEY',

  baseURL: 'https://api.lumenfall.ai/openai/v1'

});

const response = await client.images.edit({

  model: 'flux.2-dev',

  image: fs.createReadStream('source.png'),

  prompt: 'Add a starry night sky to this image',

  size: '1024x1024'

});

// { created: 1234567890, data: [{ url: "https://...", revised_prompt: "..." }] }

console.log(response.data[0].url);

from openai import OpenAI

client = OpenAI(

    api_key="YOUR_API_KEY",

    base_url="https://api.lumenfall.ai/openai/v1"

response = client.images.edit(

    model="flux.2-dev",

    image=open("source.png", "rb"),

    prompt="Add a starry night sky to this image",

    size="1024x1024"

# { created: 1234567890, data: [{ url: "https://...", revised_prompt: "..." }] }

print(response.data[0].url)

Parameter Reference

Required Supported Not available

Core Parameters

Parameter	Type	Description	Modes
`prompt`	string	Required. Edit instruction for the image	T2I Edit
`seed`	integer	Random seed for reproducibility	T2I Edit

Size & Layout

Parameter	Type	Description	Modes
`size`	string	Image dimensions as WxH pixels (e.g. "1024x1024") or aspect ratio (e.g. "16:9") WxH determines both shape and scale (aspect_ratio and resolution are ignored when size is provided). W:H format is equivalent to aspect_ratio.	T2I Edit
`aspect_ratio`	string	Aspect ratio of the output image (e.g. "16:9", "1:1") Controls shape independently of scale. Use with resolution to control both. If size is also provided, size takes precedence. Any ratio is accepted and mapped to the nearest supported value.	T2I Edit
`resolution`	string	Output resolution tier (e.g. "1K", "4K") `auto` `1K` Controls scale independently of shape. Higher tiers produce larger images and cost more. If size is also provided, size takes precedence for scale. Any tier is accepted and mapped to the nearest supported value.	T2I Edit

Output	`size`		`aspect_ratio` + `resolution`
Flexible
Auto	`"auto"`	—	Model chooses optimal dimensions
Custom 1–14142px per side	`"WxH"`	—	Any pixel dimensions within model constraints

1K 9 sizes

Output	`size`		`aspect_ratio` + `resolution`
1183 × 887	`"1183x887"`	or	`"4:3"` + `"1K"`
916 × 1145	`"916x1145"`	or	`"4:5"` + `"1K"`
1145 × 916	`"1145x916"`	or	`"5:4"` + `"1K"`
1024 × 1024	`"1024x1024"`	or	`"1:1"` + `"1K"`
887 × 1182	`"887x1182"`	or	`"3:4"` + `"1K"`
836 × 1254	`"836x1254"`	or	`"2:3"` + `"1K"`
1254 × 836	`"1254x836"`	or	`"3:2"` + `"1K"`
768 × 1365	`"768x1365"`	or	`"9:16"` + `"1K"`
1365 × 768	`"1365x768"`	or	`"16:9"` + `"1K"`

How these parameters work

size

Exact pixel dimensions

"1920x1080"

aspect_ratio

Shape only, default scale

"16:9"

resolution

Scale tier, preserves shape

"1K"

Priority when combined

size › aspect_ratio + resolution › aspect_ratio › resolution

size is most specific and always wins. aspect_ratio and resolution control shape and scale independently.

How matching works

Shape matching – we pick the closest supported ratio. Ask for 7:1 on a model with 4:1 and 8:1, you get 8:1.

Scale matching – providers use different tier formats: K tiers (0.5K 1K 2K 4K) or megapixel tiers (0.25 1). If the exact tier isn't available, you get the nearest one.

Dimension clamping – if a model has pixel limits, we clamp dimensions to fit and keep the aspect ratio intact.

Media Inputs

Parameter	Type	Description	Modes
`image`	file	Required. Input image(s) to edit Supports PNG, JPEG, WebP.	T2I Edit

Output & Format

Parameter	Type	Description	Modes
`response_format`	string	How to return the image `url` `b64_json` Default: `"url"`	T2I Edit
`output_format`	string	Output image format `png` `jpeg` `gif` `webp` `avif` Gateway converts to requested format if provider doesn't support it natively.	T2I Edit
`output_compression`	integer	Compression level for lossy formats (JPEG, WebP, AVIF)	T2I Edit
`n`	integer	Number of images to generate Default: `1` Gateway generates multiple images in parallel even if provider only supports 1.	T2I Edit

Additional Parameters

Parameter	Type	Description	Modes
`cfg_scale`	number	Classifier-free guidance scale — higher values stick more closely to the prompt	T2I Edit
`acceleration` fal	string	The acceleration level to use for the image generation. `high` `none` `regular`	T2I Edit
`disable_safety_checker` replicate	boolean	Disable safety checker for generated images.	T2I Edit
`enable_prompt_expansion` fal	boolean	If set to true, the prompt will be expanded for better results.	T2I Edit
`enable_safety_checker` fal	boolean	If set to true, the safety checker will be enabled.	T2I Edit
`go_fast` replicate	boolean	Run faster predictions with additional optimizations.	T2I Edit
`height` replicate	integer	Height of the generated image in text-to-image mode. Only used when aspect_ratio=custom. Must be a multiple of 32 (if it's not, it will be rounded to nearest multiple of 32).	T2I Edit
`num_inference_steps` fal	integer	The number of inference steps to perform.	T2I Edit
`output_quality` replicate	integer	Quality when saving the output images, from 0 to 100. 100 is best quality, 0 is lowest quality. Not relevant for .png outputs.	T2I Edit
`sync_mode` fal	boolean	If `True`, the media will be returned as a data URI and the output data won't be available in the request history.	T2I Edit
`width` replicate	integer	Width of the generated image in text-to-image mode. Only used when aspect_ratio=custom. Must be a multiple of 32 (if it's not, it will be rounded to nearest multiple of 32).	T2I Edit

Parameter Normalization

How we handle parameters across different providers

Not every provider speaks the same language. When you send a parameter, we handle it in one of four ways depending on what the model supports:

Behavior	What happens	Example
`passthrough`	Sent as-is to the provider	style, quality
`renamed`	Same value, mapped to the field name the provider expects	prompt
`converted`	Transformed to the provider's native format	size
`emulated`	Works even if the provider has no concept of it	n, response_format

Parameters we don't recognize pass straight through to the upstream API, so provider-specific options still work.

Full API Reference Authentication, endpoints, and more

FLUX.2 [dev] Benchmarks

FLUX.2 [dev] maintains a competitive position in high-end image generation with a Text-to-Image Elo of 1236, currently ranking #18 globally. This open-weights model from Black Forest Labs delivers frontier-level visual fidelity comparable to top-tier proprietary systems.

Lumenfall Arena

#20

Text-to-Image

1233 Elo

Text-to-Image Landscape

Elo vs Cost

Elo vs Speed

10 without speed data omitted.

Competition Results

Text-to-Image

Photorealism

View leaderboard

Candid Street Photography

24 models

Text-to-Image

Prompt

“A candid street photo of an elderly Japanese man repairing a red bicycle in light rain, reflections on wet pavement, shallow depth of field, 50mm lens, natural skin texture, imperfect framing, motion blur from passing cars, cinematic but realistic, no stylization.”

Text-to-Image

Portrait

View leaderboard

#14

Fantasy Warrior

21 models

Text-to-Image

Prompt

“Close portrait of a battle-worn paladin in ornate engraved plate armor, hair braided with small beads, faint scars and dirt on the skin, warm torchlight reflecting off metal, shallow depth of field, bokeh sparks, lifelike eyes, highly detailed texture on leather straps and cloth underlayer.”

Uncategorized

#14

Geometric Composition

22 models

Text-to-Image

Prompt

“A glass cube on a wooden table. Inside the cube is a small blue sphere. On top of the cube sits a red book. A green plant is behind the cube, partially visible through the glass. Soft window light from the left.”

#22

Adorable Baby Animals in Sunny Meadow

25 models

Text-to-Image

Prompt

“Hyper-photorealistic scene of fluffy baby animals—a golden retriever puppy, tabby kitten, baby bunny, and red fox kit—with big expressive eyes and ultra-detailed soft fur, playfully chasing butterflies and tumbling together in a lush wildflower meadow, warm golden sunrise light with god rays and dew sparkles, joyful wholesome vibe, 8K masterpiece.”

#20

Heroic Super Hero Portrait

21 models

Text-to-Image

Prompt

“Hyper-photorealistic full-body portrait of a female superhero standing triumphantly on a New York skyscraper rooftop at golden sunset, wearing a classic modest superhero costume with flowing cape, chest emblem, gloves, and boots in red and blue colors, practical design, short hair, strong determined heroic expression looking into the distance, powerful confident stance with hands on hips and cape billowing dramatically in the wind, detailed urban cityscape background, warm natural sunlight with sharp shadows and fabric highlights, ultra-sharp textures on suit, hair, and concrete, 8K masterpiece, empowering family-friendly style.”

Fantasy Warrior

14 models

Text-to-Image

Prompt

3 attempts – showing best result

Top Matchups

See how FLUX.2 [dev] performs head-to-head against other AI models, ranked by community votes in blind comparisons.

vs Nano Banana 2

Challenge: Fantasy Warrior

0% W · 100% L

vs Nano Banana

Challenge: Heroic Super Hero Portrait

0% W · 100% L

Use Cases

See all Use Cases

The model performs strongest in portrait generation where it holds rank #11 with a 42.9% win rate, though its photorealism ranking sits at #15. It excels at text-to-image synthesis and complex image editing tasks where anatomical accuracy and prompt adherence are critical.

#12 of 21

Portrait

Gen · 42.3% win rate

#14 of 23

Photorealism

Gen · 42.1% win rate

Gallery

View all 10 images

FLUX.2 [dev] FAQ

How much does FLUX.2 [dev] cost?

FLUX.2 [dev] starts at $0.012 per image through Lumenfall. Pricing varies by provider. Lumenfall does not add any markup to provider pricing.

How do I use FLUX.2 [dev] via API?

You can use FLUX.2 [dev] through Lumenfall's OpenAI-compatible API. Send requests to the unified endpoint with model ID "flux.2-dev". Code examples are available in Python, JavaScript, and cURL.

Which providers offer FLUX.2 [dev]?

FLUX.2 [dev] is available through fal.ai and Replicate on Lumenfall. Lumenfall automatically routes requests to the best available provider.

What is the maximum resolution for FLUX.2 [dev]?

FLUX.2 [dev] supports images up to 2048x2048 resolution.

Overview

FLUX.2 [dev] is an open-weights image generation model developed by Black Forest Labs, designed to offer frontier-level performance for non-commercial applications. It serves as an intermediate iteration between high-speed distilled models and large-scale professional versions, balancing computational efficiency with high visual fidelity. The model is specifically engineered to handle complex text-to-image prompts through a refined Rectified Flow architecture.

Strengths

High Text Rendering Accuracy: The model demonstrates significant improvements in rendering legible, correctly spelled text within generated images, even in complex layouts or unconventional fonts.
Instruction Adherence: It excels at following multi-part prompts that specify spatial relationships, color palettes, and specific lighting conditions without losing detail in the background.
Anatomical Realism: Compared to previous iterations in the FLUX family, this version shows increased stability in generating human anatomy, particularly regarding hands, limb articulation, and skin textures.
Compositional Diversity: The model is less prone to “canonical” centering, allowing for more dynamic framing and varied perspectives based on descriptive text.

Limitations

Non-Commercial Licensing: Unlike the “schnell” variants or standard open-source models, FLUX.2 [dev] is restricted to non-commercial use, which limits its application in production environments or for-profit products.
Hardware Requirements: While designed for local deployment, the model still requires significant VRAM to run at full precision, making it less accessible for entry-level consumer GPUs without quantization.
Inference Latency: It prioritizes output quality over generation speed, meaning it is noticeably slower than distilled 4-step models.

Technical Background

FLUX.2 [dev] is built on a Rectified Flow-based transformer architecture, which improves upon traditional diffusion methods by straightening the trajectory from noise to image. This approach allows for more efficient sampling and better alignment between the text encoder and the visual output. The training process leverages a massive-scale dataset designed to enhance the model’s understanding of complex semantics and nuanced visual concepts.

Best For

This model is best suited for visual researchers, creative hobbyists, and developers prototyping new image generation workflows who require high-quality visual outputs without the constraints of a closed API. It is particularly useful for projects requiring precise typography or complex scene composition. FLUX.2 [dev] is available for experimentation and integration through Lumenfall’s unified API and interactive playground, allowing you to compare its performance against other models in its class.

Try FLUX.2 [dev] in Playground

Generate images with custom prompts — no API key needed.

Open Playground