FLUX.2 [dev] LoRA
FLUX.2 [dev] LoRA is an image generation model from Black Forest Labs that supports LoRA adapters and reference image inputs.
LoRA-powered image generation with reference inputs
FLUX.2 [dev] LoRA is an image generation model developed by Black Forest Labs and made available through the Wavespeed provider on MindStudio. It is a variant of the FLUX.2 [dev] base model that adds native support for LoRA (Low-Rank Adaptation) weights, allowing users to apply fine-tuned style or subject adapters on top of the base model at inference time. The model accepts reference images alongside text prompts, enabling guided generation that draws from visual inputs. It was released in November 2025 and is priced at $0.018 per generated image.
This model is well suited for workflows that require stylistic consistency or subject fidelity across generated images, particularly when paired with custom LoRA checkpoints trained on specific aesthetics, characters, or products. Users can specify output dimensions via width and height inputs and control reproducibility through a seed parameter. The reference image input allows the model to incorporate visual context from existing images, making it useful for tasks like style transfer, character consistency, and brand-aligned content generation. Its LoRA support means it can be adapted to a wide range of specialized use cases without retraining the full model.
What FLUX.2 [dev] LoRA supports
LoRA Adapter Support
Accepts one or more LoRA weight files at inference time, allowing fine-tuned styles or subjects to be applied on top of the base FLUX.2 [dev] model without full retraining.
Reference Image Input
Takes an array of reference image URLs as input to guide generation, enabling visual context from existing images to influence the output.
Custom Output Dimensions
Accepts explicit width and height values as numeric inputs, giving users direct control over the pixel dimensions of generated images.
Seed Control
Supports a seed parameter that makes generation reproducible, allowing the same prompt and settings to produce consistent results across runs.
Text-to-Image Generation
Generates images from text prompts using the FLUX.2 [dev] architecture, producing outputs at $0.018 per image via the Wavespeed provider.
Ready to build with FLUX.2 [dev] LoRA?
Get Started FreeCommon questions about FLUX.2 [dev] LoRA
How much does it cost to generate an image with FLUX.2 [dev] LoRA?
Each generated image costs $0.018, as listed in the model's published pricing on MindStudio.
What is the context window for this model?
The model has a context window of 10,000 tokens, which applies to the text prompt input.
Can I use my own LoRA weights with this model?
Yes. The model includes a dedicated LoRAs input field that accepts LoRA adapter weights, allowing you to apply custom fine-tuned adapters at inference time.
Does this model support reference images?
Yes. The model accepts an array of image URLs via the Reference Images input, which can be used to guide the visual output of generation.
When was FLUX.2 [dev] LoRA released?
The model was released in November 2025 and was added to MindStudio on January 12, 2026.
Documentation & links
Parameters & options
If you want to edit an existing image, provide the URL(s) or variables
Up to 3 LoRAs.
A specific value that is used to guide the 'randomness' of the generation.
Explore similar models
Start building with FLUX.2 [dev] LoRA
No API keys required. Create AI-powered workflows with FLUX.2 [dev] LoRA in minutes — free.