Flux.2 [flex]
The tunable Flux.2 tier — trade denoising steps against speed per generation.
All modelsOverview
Flux.2 [flex] is an image generation and editing model developed by Black Forest Labs. While [Flux.2 pro] focuses on cinematic photorealism, this specialized model is purpose-built for text-heavy workflows, handling complex typography, infographics, and UI prototyping. It preserves fine details in dense scenes and allows users to adjust inference steps to balance speed and quality, making it a practical choice for precise brand design and multi-reference compositing.
Specifications · Text + image to image
- Input mode
- Text + image to image
- Accepts
- reference image (up to 8)
- Aspect ratios
- 16:9, 9:16, 4:3, 3:4, 1:1
Pricing
Build with this model: the Flux.2 [flex] API on the Hedra Developer Platform.
Flux.2 [flex] API →Real output · generated with Flux.2 [flex]
Best of Flux.2 [flex]
Prompting
Prompt tips
- Structure prompts as JSON: Use JSON formatting with specific keys like
scene,subjects,style,lighting, andcamerato force strict compositional control. - Use positive constraints: Instead of writing "no messy background," use affirmative descriptions like "clean, solid white background."
- Specify HEX color codes: Include exact HEX codes in your text prompt to maintain strict brand color guidelines in product shots or UI designs.
- Stack reference images: Upload multiple reference images in a single generation to combine a specific subject's likeness with a completely different artistic style.
About the model
Questions, answered
What types of generations is Flux.2 [flex] best suited for?
Flux.2 [flex] is specialized for text-heavy design, typography, and UI prototyping. While [Flux.2 pro] focuses on pure photorealism, the flex variant is optimized to render clean text, logos, and complex layouts like infographics. It also supports multi-reference editing with up to 10 reference images, making it highly effective for maintaining character or product consistency across various scenes and compositions.
When was Flux.2 [flex] released, and how does it fit into the Flux family?
Black Forest Labs introduced the Flux.2 family on November 25, 2025, succeeding earlier models like Flux 1.1 Pro and Flux 1.1 Ultra. The lineup includes the flagship [Flux.2 pro], the high-fidelity [Flux.2 max], the fast [Flux.2 klein] 9B, and this configurable flex model. Unlike its siblings that lock down parameters for consistency, Flux.2 [flex] exposes inference steps and guidance scale, allowing developers to manually adjust the speed-to-quality ratio.
How can I get the best results from Flux.2 [flex]?
Take advantage of its exposed parameters by adjusting the guidance scale (CFG) and step count to find your preferred balance of speed and detail. For precise layouts, the model responds well to structured JSON prompting and exact hex color codes. When doing multi-image compositing, supply up to 10 reference images to blend characters and styles seamlessly. For detailed prompt formatting advice, review the official Black Forest Labs blog.