Bria's Fibo 2.0 Brings Structured Prompts to Open Image AI

Bria has released Fibo 2.0-dev, a new version of its text-to-image and image-editing model, published on Hugging Face. The model is aimed at users who want finer control over generation, supporting structured prompts and multiple reference images.

Fibo 2.0 handles two related jobs: generating images from text and editing existing ones. The structured-prompt approach is the differentiator here — rather than relying solely on free-form natural language, it lets users specify components of a scene in a more organized way, which tends to produce more predictable and repeatable results.

Why it matters

Most open image models compete on raw fidelity, but controllability is often the harder problem for production use. Multi-reference support and structured prompting are the kind of features that matter when an image needs to match a brand, a layout, or a prior output rather than simply look good in isolation.

• Text-to-image generation plus image editing in one model
• Structured prompts for more controllable results
• Multi-reference inputs to guide outputs

One important caveat: Fibo 2.0-dev ships under a CC BY-NC 4.0 license, which restricts commercial use. That makes it well suited to research and experimentation, but teams planning to build products on top of it will need to check Bria's commercial terms before deploying.