How to run Emu3-Gen locally

BAAI · Next token any to any generation · 8B params · Apache-2.0 (commercial OK)

An 8B transformer trained from scratch that tokenizes text, images and video into one discrete space and generates images and short video purely by next-token prediction, with no diffusion component.

What it costs to run — $0 markup

On your own machine — $0. Runs free if you have about 18 GB of VRAM (~16-18GB in BF16 for the 8B generation model on a 24GB card; quantization brings it lower at some quality cost.), via Transformers (Emu3Processor).
Rent a GPU — from $0.26/hr. Fits RTX 3090 24GB at the direct vendor price ($0 markup) — pay only for the minutes you generate.
Download the weights — free. Open weights at BAAI/Emu3-Gen.

Note: generative-media models are billed per image / per second / per minute on hosted services — not per token. Running locally or on your own rented GPU is usually far cheaper and keeps your data on your machine.

Key facts

Does	Text → image, Text → video, Video prediction, Visual understanding
VRAM to run	~18 GB (~16-18GB in BF16 for the 8B generation model on a 24GB card; quantization brings it lower at some quality cost.)
Download	~17 GB
Parameters	8B
License	Apache-2.0 (commercial use OK)
Run with	Transformers (Emu3Processor)

Get Emu3-Gen on Hugging Face →

More multimodal / omni models

Qwen2.5-Omni-7B — Omni understanding + speech generation, ~24 GB VRAM
Janus-Pro-7B — Unified understanding + image gen, ~16 GB VRAM
OmniGen2 — Any to image (gen + edit + in context), ~17 GB VRAM
MiniCPM-o 2.6 — Omni understanding + speech generation, ~9.0 GB VRAM
BAGEL-7B-MoT — Unified understanding + image gen + edit, ~24 GB VRAM
Emu3.5 — Any to any world model (gen + edit), ~48 GB VRAM

Browse: all media models · chat / LLM models