text-to-image
Idle
Your request will cost $0.04 per run.
For $1 you can run this model approximately 25 times.
FLUX1.1 [pro] delivers image generation speeds up to six times faster than the previous FLUX.1 [pro], while also enhancing visual fidelity, prompt accuracy, and output variety.
With reduced latency and quicker generation, FLUX1.1 [pro] streamlines your creative process, striking a strong balance between output quality and performance.
Codenamed “blueberry” during development, FLUX1.1 [pro] was evaluated in the Artificial Analysis image arena—one of the leading benchmarks for text-to-image models—and outperformed all competitors, earning the top Elo rating.
The FLUX.1 series utilizes a hybrid design combining multimodal and parallel diffusion transformer modules, scaled to 12 billion parameters. By leveraging flow matching—a straightforward yet powerful approach to generative model training that generalizes diffusion—these models surpass previous diffusion-based methods. Additional improvements include rotary positional embeddings and parallel attention, boosting both model capability and hardware efficiency.