Model reference · open weights
Fibo-2.0 is an open-weight image model from Comfy-Org. Fibo-2.0 (BF16) weighs 65.8 GB; the smallest configuration that runs it is H100 80 GB.
For self-hosting, compare this build's requirements below with the GPU you plan to rent.
Summary of the Comfy-Org/Fibo-2.0 model card, 2026-10-28
What it is
| Released by | Comfy-Org |
|---|---|
| Released | 2026-10-20 |
| VRAM | 65.8 GB for the weights |
What it runs on
| Card | One image | Memory |
|---|---|---|
| RTX 3060 12 GB … L40S 48 GB | does not fit | |
| A100 80 GB | tight | 78.2 GB |
| H100 80 GB | tight | 78.1 GB |
| RTX PRO 6000 Blackwell 96 GB | fits | 93.8 GB |
| DGX Spark (GB10) 128 GB unified | fits | 107 GB |
| H200 141 GB | fits | 138 GB |
| B200 180 GB | fits | 176 GB |
From the model card
Repackaged model files for ComfyUI.
Original model repository:
Place the files in the following folders:
📂 ComfyUI/
├── 📂 models/
│ ├── 📂 diffusion_models/
│ │ ├── fibo2_dev_bf16.safetensors
│ │ ├── fibo2_dev_int8_convrot.safetensors
│ │ ├── fibo2_edit_turbo_bf16.safetensors
│ │ └── fibo2_edit_turbo_int8_convrot.safetensors
│ ├── 📂 text_encoders/
│ │ ├── qwen3vl_4b_bf16.safetensors
│ │ └── qwen3vl_4b_int8_convrot.safetensors
│ ├── 📂 vae/
│ │ └── flux2-vae.safetensors
The model weights are © BRIA AI and licensed under Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0). You may use them for non-commercial purposes only. Commercial use requires a separate agreement with BRIA AI: https://platform.bria.ai/licensing/pricing?model=fibo-2
The weights are provided as is, without warranties of any kind. See Section 5 of the license (Disclaimer of Warranties and Limitation of Liability).
bf16: the diffusers transformer shards merged into one file, with keys renamed to ComfyUI's layout. Attention to_q/to_k/to_v weights are concatenated into qkv, and the cross-attention and Perceiver to_k/to_v into to_kv. Weight values are unchanged (bf16).int8_convrot.safetensors: the same weights quantized to int8 (Hadamard rotation, per-channel scales), with the adaLN modulation layers kept in bf16. Outputs differ slightly from the original.w1/w3 fused into w13 and the FFN hidden width zero-padded to 9728@article{gutflaish2025generating,
title={Generating an Image From 1,000 Words: Enhancing Text-To-Image With Structured Captions},
author={Gutflaish, Eyal and Kachlon, Eliran and Zisman, Hezi and Hacham, Tal and Sarid, Nimrod and Visheratin, Alexander and Huberman, Saar and Davidi, Gal and Bukchin, Guy and Goldberg, Kfir and others},
journal={arXiv preprint arXiv:2511.06876},
year={2025}
}Quoted from the model card on Hugging Face. The full card is behind the Hugging Face link above.
How it works
Running it yourself
Rent a machine by the hour. ComfyUI is installed on it. Open ComfyUI through the tunnel and load the workflow from the model's card on Hugging Face; choose this model's file in its loader.
# on your rented machine (the ssh line is on its page in the console)
# get REPO FILE FOLDER: one file into /workspace/models/FOLDER, where ComfyUI loads it from
get() { hf download "$1" "$2" --local-dir /workspace/hf-files && mkdir -p "/workspace/models/$3" && mv "/workspace/hf-files/$2" "/workspace/models/$3/$4"; }
# the model (8.1 GB)
get Comfy-Org/Fibo-2.0 diffusion_models/fibo2_dev_int8_convrot.safetensors diffusion_models
get Comfy-Org/Fibo-2.0 vae/flux2-vae.safetensors vae
start-comfyui# on your computer, in a second terminal: ComfyUI in your browser at http://localhost:8188
# HOST, PORT and BOOKING are your machine's, from its page in the console
ssh -o HostKeyAlias=axforge-BOOKING -o StrictHostKeyChecking=accept-new -L 8188:localhost:8188 dev@HOST -p PORT