Skip to content
ModelListedExternal
Download

HiDream O1 Image BF16

This is the BF16 conversion of HiDream-O1-Image for use with ComfyUI. Weights have been cast to bfloat16 for a balance of precision and memory efficiency.

Sourcehuggingface.co/drbaph/HiDream-O1-Image-BF16/tree/b9d73eca0d40135d62f03b707d7afa95b2cbe571
KindText modelVersionmodelBgrade

Model source

Kind
Text model
File size
16.4 GB
Version
vmodel
Updated
Oct 9, 2026
Downloads
1.2K
Source
Hugging Face

Source description

HiDream-O1-Image — BF16 (ComfyUI)

This is the BF16 conversion of HiDream-O1-Image for use with ComfyUI. Weights have been cast to bfloat16 for a balance of precision and memory efficiency.

image

Custom ComfyUI Node: Saganaki22/HiDream_O1-ComfyUI

Screenshot 2026-05-10 005045

VRAM Requirements

PrecisionApproximate VRAM
BF16 (this repo)17 – 20 GB
FP1617 – 20 GB
FP8 Mixed~10 GB

A GPU with at least 20 GB VRAM is recommended for comfortable use at full 2048 × 2048 resolution. 24 GB cards (RTX 3090/4090, A5000, etc.) will have no issues.

Quick Start — ComfyUI

1. Install the Custom Node

cd ComfyUI/custom_nodes
git clone https://github.com/Saganaki22/HiDream_O1-ComfyUI
pip install -r HiDream_O1-ComfyUI/requirements.txt

Or install via ComfyUI Manager by searching for HiDream O1.

2. Download the Weights

huggingface-cli download drbaph/HiDream-O1-Image-BF16 \
    --local-dir ComfyUI/models/diffusion_models/HiDream-O1-Image-bf16

3. Load in ComfyUI

Open ComfyUI and use the workflow provided in the custom node repository. Point the model loader to HiDream-O1-Image-bf16.

About HiDream-O1-Image

HiDream-O1-Image is a natively unified image generative foundation model built on a Pixel-level Unified Transformer (UiT) — no external VAEs, no disjoint text encoders. It encodes raw pixels, text, and task-specific conditions in a single shared token space, supporting:

  • Text-to-image generation up to 2,048 × 2,048
  • Instruction-based image editing
  • Subject-driven personalization (multi-reference IP)
  • Long-text and multilingual text rendering

At only 9B parameters it matches or exceeds much larger open-source DiTs and leading closed-source models. It debuted at #8 in the Artificial Analysis Text to Image Arena (2026-05-05).

Key Features

  • 🧬 Pixel-Level Unified Transformer — end-to-end on raw pixels, no VAE, no disjoint text encoder
  • 🎨 One Model, Many Tasks — T2I, editing, personalization, storyboard generation
  • 🧠 Reasoning-Driven Prompt Agent — built-in "thinking" agent that resolves layout and rendering before generation
  • 🖼️ Native High Resolution — direct synthesis up to 2,048 × 2,048
  • ⚡ 9B Parameters — performance parity with models many times larger

Model Variants

RepoPrecisionVRAMInference Steps
drbaph/HiDream-O1-Image-BF16 (this repo)BF1617–20 GB50
drbaph/HiDream-O1-Image-FP16FP1617–20 GB50
drbaph/HiDream-O1-Image-FP8FP8 Mixed~10 GB50
HiDream-ai/HiDream-O1-ImageOriginal—50
HiDream-ai/HiDream-O1-Image-DevOriginal Dev—28

Benchmark Results (from original model)

GenEval (compositional generation) — HiDream-O1-Image scores 0.90 overall at 9B params, second only to the 200B+ Pro variant and ahead of GPT Image 2 (0.89).

DPG-Bench (dense prompt alignment) — Overall score 89.83, ranking second behind the Pro variant.

HPSv3 (human preference) — Overall score 10.37, outperforming GPT Image 2 (10.21) and Nano Banana 2.0 (10.01).

License

The original HiDream-O1-Image model and code are released under the MIT License. This BF16 conversion inherits the same license.

Links

Sources

1 source
Hugging FacePrimaryVersion b9d73eca0d40135d62f03b707d7afa95b2cbe571
huggingface.co/drbaph/HiDream-O1-Image-BF16/tree/b9d73eca0d40135d62f03b707d7afa95b2cbe571

Verified Oct 9

Model artifacts

1 artifact

model.safetensors

safetensors · 16.4 GB · SHA-256 2c0cfb6d8372…5ef1 · Hugging Face

Download

Source excerpts

3 excerpts
ReadmeJul 31huggingface.co/drbaph/HiDream-O1-Image-BF16

--- license: mit pipeline_tag: image-text-to-image library_name: diffusers tags: - comfyui - text-to-image - image-editing - bf16 language: - en - zh --- # HiDream-O1-Image — BF16 (ComfyUI) This is the **BF16** conversion of [HiDream-O1-Image](https://huggingface.co/HiDream-ai/HiDream-O1-Image) for use with **ComfyUI**. Weights have been cast to `bfloat16` for a balance of precision and memory efficiency. ![image](https://cdn-uploads.huggingface.co/production/uploads/63473b59e5c0717e6737b872/MRw4jv7Zcc48EchtsXu1b.png) > **Custom ComfyUI Node:** [Saganaki22/HiDream_O1-ComfyUI](https://github.com/Saganaki22/HiDream_O1-ComfyUI) ![Screenshot 2026-05-10 005045](https://cdn-uploads.huggingface.co/production/uploads/63473b59e5c0717e6737b872/iApQk_Ms8DPmMQroZ2Ge-.png) --- ## VRAM Requirements | Precision | Approximate VRAM | | :--- | :---: | | BF16 (this repo) | **17 – 20 GB** | | FP16 | 17 – 20 GB | | FP8 Mixed | ~10 GB | A GPU with at least **20 GB VRAM** is recommended for comfortable use at full 2048 × 2048 resolution. 24 GB cards (RTX 3090/4090, A5000, etc.) will have no issues. --- ## Quick Start — ComfyUI ### 1. Install the Custom Node ```bash cd ComfyUI/custom_nodes git clone https://github.com/Saganaki22/HiDream_O1-ComfyUI pip install -r HiDream_O1-ComfyUI/requirements.txt ``` Or install via **ComfyUI Manager** by searching for `HiDream O1`. ### 2. Download the Weights ```bash huggingface-cli download drbaph/HiDream-O1-Image-BF16 \ --local-dir ComfyUI/models/diffusion_models/HiDream-O1-Image-bf16 ``` ### 3. Load in ComfyUI Open ComfyUI and use the workflow provided in the custom node repository. Point the model loader to `HiDream-O1-Image-bf16`. --- ## About HiDream-O1-Image HiDream-O1-Image is a natively unified image generative foundation model built on a **Pixel-level Unified Transformer (UiT)** — no external VAEs, no disjoint text encoders. It encodes raw pixels, text, and task-specific conditions in a single shared token space, supporting: - **Text-to-image generation** up to 2,048 × 2,048 - **Instruction-based image editing** - **Subject-driven personalization** (multi-reference IP) - **Long-text and multilingual text rendering** At only 9B parameters it matches or exceeds much larger open-source DiTs and leading closed-source models. It debuted at **#8 in the Artificial Analysis Text to Image Arena** (2026-05-05). --- ##...

NoteOct 9huggingface.co/drbaph/HiDream-O1-Image-BF16

Source context: 1232 downloads · 2 likes · Pipeline image-text-to-image · Library diffusers · Repo drbaph/HiDream-O1-Image-BF16

NoteJul 31huggingface.co/drbaph/HiDream-O1-Image-BF16

Source context: 208 downloads · 2 likes · Pipeline image-text-to-image · Library diffusers · Repo drbaph/HiDream-O1-Image-BF16