More image detail does not automatically mean more control: short prompts leave models to invent many visual choices. This model tackles that gap by treating a scene as a structured blueprint, so edits can target attributes such as lighting, composition, or objects while leaving other details stable. Its main trade-off is access: the released weights are restricted to non-commercial use.
AI Model2026
Fibo 2.0 Dev
Generates and edits images from structured prompts, aiming to preserve details you did not ask to change. Supports masked edits and multiple reference images; the weights are for non-commercial use, with commercial use requiring a separate agreement.
Introduction
Information
- Websitehuggingface.co
- OrganizationsBRIA AI
- Published date2026/10/18
Categories
More Items
AI Model2026
Turns natural-language descriptions into runnable Three.js scene code, exploring how a small language model can generate structured interactive 3D environments as well as general-purpose code.
AI Model2026
Translates incoming Chinese or English speech incrementally, aiming to keep lag low without losing meaning. This FP8 checkpoint returns translated text; speech synthesis and voice cloning require a separately configured TTS service.
AI Model2026
Runs local agents, tool calls, and structured-output tasks with a 256K-token context window. At 4-bit precision it uses about 1 GB of memory, and it can be fine-tuned for specialized domains on a single GPU.