FIBO Models by BRIA

BRIACC-BY-NC-4.0Open weights

1 model2025

Last refreshed 2026-05-19. Next refresh: weekly.

Details

ResearcherBRIA

LicenseCC-BY-NC-4.0

Commercial useCommercial use: non-commercial

Models1

Released2025

Capabilities

MultimodalAll models

Links

Website HuggingFace

About

Bria AI's FIBO (Foundation Image Base with Order) model family — the first open-source JSON-native text-to-image model. Trained exclusively on licensed data, FIBO uses structured JSON prompts (~1,000 words each) encoding 100+ visual attributes for deterministic, brand-safe image generation.

Decision facts

Best fit: image generationvision and multimodal workcoding
Capability starting point: FIBO with multimodal inputs
Lowest tracked input: Not tracked
Closest related family: Bria

Current Variants

Use-when guidance is based on each model's tracked capabilities, context window, release date, and replacement status.

1 in view

FIBOCurrent

Use when the workload needs image generation, 8B parameters, and multimodal inputs.

2025-11image generation8B parametersmultimodal inputs

Current FIBO variants with use-when guidance and lifecycle status
Model	Use when	Released	Signals	Status
FIBO	Use when the workload needs image generation, 8B parameters, and multimodal inputs.	2025-11	image generation8B parametersmultimodal inputs	Current

Release Timeline

1 release group

2025-11

1 current

FIBO

image generation8B parametersmultimodal inputs

Current

Specifications(1 models)

FIBO model specifications comparison
Model	Released	Parameters	Multimodal
FIBO	2025-11	8B	Yes

Models(1)