How to run Qwen2.5-Omni-7B locally

Alibaba Qwen · Omni understanding + speech generation · 11B params · Apache-2.0 (commercial OK)

End-to-end Thinker-Talker model that takes text, image, audio and video in and streams back both text and natural speech, making it the flagship open omni model that actually talks. Note the 3B sibling exists but ships under a non-commercial qwen-research license.

What it costs to run — $0 markup

On your own machine — $0. Runs free if you have about 24 GB of VRAM (~31GB in BF16; fits a single 24GB card with flash-attention + reduced context, or use the official AWQ 4-bit build for ~12GB. CPU offload possible but slow for the streaming Talker.), via Transformers (qwen-omni-utils).
Rent a GPU — from $0.49/hr. Fits RTX A6000 48GB at the direct vendor price ($0 markup) — pay only for the minutes you generate.
Download the weights — free. Open weights at Qwen/Qwen2.5-Omni-7B.

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	Visual understanding, Audio understanding, Video understanding, Text → speech, Speech → speech, Any → text
VRAM to run	~24 GB (~31GB in BF16; fits a single 24GB card with flash-attention + reduced context, or use the official AWQ 4-bit build for ~12GB. CPU offload possible but slow for the streaming Talker.)
Download	~31 GB
Parameters	11B
License	Apache-2.0 (commercial use OK)
Run with	Transformers (qwen-omni-utils)

Get Qwen2.5-Omni-7B on Hugging Face →

More multimodal / omni models

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
Emu3-Gen — Next token any to any generation, ~18 GB VRAM

Browse: all media models · chat / LLM models