Skip to content

EthenEthenEthen

LoginPlatformTry Ethen

Open Source Model Profile · krea

Krea-2-Turbo

Krea-2-Turbo is a 12.82B-parameter text-to-image diffusion model from Krea. Its model card documents a distilled Turbo checkpoint with 8-step Krea2Pipeline inference and Community License terms.

Publisher
krea
Task
text-to-image
Model type
Unknown
License
krea-2-community-license
Library
diffusers
Publication status
Accepted · not indexed

Model overview

Krea-2-Turbo is published by Krea as a text-to-image model with 12,820,073,036 Safetensors parameters, or about 12.82B. According to the model card, it is the post-trained Turbo checkpoint of the Krea 2 Diffusion Transformer family from Krea.ai, Inc., distilled with extra fine-tuning over the Raw base. Captured metadata records diffusers library support and an other license value.

Recorded capabilities

Distilled Turbo checkpoint

The card identifies Turbo as the post-trained release with additional fine-tuning and distillation over the Raw base.

Documented 8-step inference

The card shows 8-step generation via the official codebase, diffusers Krea2Pipeline, and an SGLang path.

Explicit license and safety terms

The card sets out-of-scope uses, deployer filtering duties, and user responsibility for outputs under the Krea 2 Community License.

Use cases in the source record

  • Text-to-image generation from natural-language prompts at resolutions up to 2048x2048 through diffusers, SGLang, or the official codebase.
  • Application development where the team performs its own safety testing and implements the filtering safeguards the license requires.

Limitations and unknowns

  • No evaluation results were extracted from this record.
  • Provider state is historical snapshot data, not independently refreshed current availability.
  • The card notes the model is not intended to provide factual information and prompt following varies with style and phrasing.

Source and provenance

Source: krea/Krea-2-Turbo

Captured: Unknown. Processed: 2026-09-07T19:34:49.234084+00:00.

Krea 2 Text-to-Image Model Inference with the official codebase Setup the official Krea 2 codebase Download turbo.safetensors in this repo export OSS_TURBO=<path-to-turbo.safetensors> Run inference: uv run inference.py "a fox walking in the snow" \ --checkpoint oss_turbo --steps 8 --cfg 0.0 --mu 1.15 --width 2048 --height 2048 Inference with diffusers Install diffusers from source (for Krea2Pipeline ): pip install git+https://github.com/huggingface/diffusers.git import torch from diffusers import Krea2Pipeline pipe = Krea2Pipeline.from_pretrained( "krea/Krea-2-Turbo" , torch_dtype=torch.bfloat16).to( "cuda" ) image = pipe( "a fox in…

F001F002F003F004F005F006F007F008F009F010F012F013F016F017F018F019