Krea 2
Frontierby Krea
Krea 2 (raw base) is a 12B-parameter DiT text-to-image model from Krea.ai, released as a foundation for fine-tuning and creative/commercial use. The raw base favors aesthetic flexibility over baked-in style; a Turbo distilled variant is also available.
VRAM requirements, GPU fit, and setup notes for Krea 2, including 8GB/12GB fit guidance where relevant. Recommended runtimes: ComfyUI and Diffusers support. Best download size: ~12.0 GB at FP8.
- 12B DiT text-to-image base from Krea.ai
- Raw base designed as a foundation for fine-tuning and LoRA training
- Turbo distilled and quantized (FP8/GGUF) community builds available
- Strong aesthetic quality and prompt adherence
Your hardware
Detecting...
Image Quality Benchmarks
Measured quality metrics for Krea 2 outputs.
How often humans prefer this model's output (0-100%)
Visual quality and composition rating (5-9 scale)
VRAM Requirements by Resolution and Precision
Compare which GPUs can run Krea 2 at different precisions. FP8 uses less memory than FP16 when available, and the grade shows how comfortably each GPU handles the workload.
FP16 (full precision)
| Resolution | VRAM Required | RTX 4090 24GB | RTX 3060 12GB | RTX 4060 8GB | MacBook Pro M4 Pro 24GB |
|---|---|---|---|---|---|
| 512×512 | 36.0 GB | F | F | F | F |
| 768×768 | 36.2 GB | F | F | F | F |
| 1024×1024 | 36.6 GB | F | F | F | F |
FP8 (~40% less VRAM)
| Resolution | VRAM Required | RTX 4090 24GB | RTX 3060 12GB | RTX 4060 8GB | MacBook Pro M4 Pro 24GB |
|---|---|---|---|---|---|
| 512×512 | 18.4 GB | S | F | F | B |
| 768×768 | 18.7 GB | S | F | F | B |
| 1024×1024 | 19.1 GB | S | F | F | B |
Optimization Tips
Turbo / LCM distillation
Use distilled scheduler at 4-8 steps for faster iteration
Run with Python
from diffusers import DiffusionPipeline
import torch
pipe = DiffusionPipeline.from_pretrained(
"krea/Krea-2-Raw",
torch_dtype=torch.float16
)
pipe.to("cuda")
image = pipe(
prompt="your prompt here",
num_inference_steps=52,
guidance_scale=4.5,
height=1024,
width=1024,
).images[0]
image.save("output.png")Get started
Setup instructions for running Krea 2 locally
1. Download the model
Get the checkpoint from HuggingFace
2. Place in:
ComfyUI/models/checkpoints/3. Launch ComfyUI
python main.pyMemory Breakdown
VRAM allocation at 1024×1024 on RTX 4090 24GB (24 GB)
Estimated Generation Time
Time per image at 1024×1024, 28 steps, FP16.
Available Formats, Downloads & Setup
Download Krea 2 in the precision that matches your GPU. Lower precision usually means less VRAM pressure, while higher precision keeps more quality.
| Format | Präzision | Größe | Anbieter | |
|---|---|---|---|---|
| Offizielle Gewichte | ||||
| safetensorsEmpfohlen | FP16 | 24.0 GB | official | Herunterladen |
| Community-Konvertierungen | ||||
| safetensorsCommunity | FP8 | 12.0 GB | community-fp8 | Herunterladen |
LoRA Ecosystem
Growing EcosystemGrowing LoRA ecosystem from Krea and the community (style, realism, and detail LoRAs).
Related Workflows
You might also like
Frequently asked questions
FAQ — Krea 2 VRAM, Runtimes & Fit
How much VRAM does Krea 2 need?
Krea 2 (12B parameters) requires approximately 36.6 GB of VRAM at FP16 precision for standard 1024×1024 image generation. If you want a lighter setup, lower precisions like FP8 can reduce memory use when available.
Can I run Krea 2 on an 8GB GPU?
Krea 2 usually needs more than 8GB for comfortable local use. Check the VRAM table above for the exact resolution and precision trade-off.
Does Krea 2 work in ComfyUI and Diffusers?
Krea 2 is marked for ComfyUI and Diffusers support in our catalog, so those are the runtimes we recommend first for local setup. If your workflow uses another front end, check the model's available formats and workflow notes above before downloading.
Can I run Krea 2 on RTX 4090?
Krea 2 is too large for the RTX 4090's 24 GB VRAM at FP16. Consider using FP8 precision or a GPU with more VRAM.
Does Krea 2 support ControlNet?
There are currently no known ControlNet adapters for Krea 2. Check Hugging Face and Civitai for community-contributed adapters.
Does Krea 2 have LoRA support?
Growing LoRA ecosystem from Krea and the community (style, realism, and detail LoRAs). The LoRA ecosystem for Krea 2 is rated as "moderate". Each LoRA adds roughly 0.3 GB of extra VRAM.
How fast is Krea 2?
On a reference GPU (RTX 4090 24GB), Krea 2 generates a 1024×1024 image in approximately ~5.4s at FP16 with 28 inference steps. Faster GPUs with higher memory bandwidth will produce images more quickly.
About Krea 2
See also