This model will generate THE best penis that I've seen for a text to image model at the time of this writing (September 2026). You can prompt vein amount, size, and other anatomy as done in my showcase images. I have an uncut version that came out great <a target="_blank" rel="ugc" href="https://civarchive.com/models/2742807">here</a>.The body and pubic hair is the best I've seen and it even handles dark wet hair, which is usually a fail (I'm talking to you Qwen Image and your squiggly lines for leg hair)When creating the showcase images, it is the first time for any of my models where I would have been happy posting any and all of them - not a single backwards dick head, double shaft, or originating out of the butt, etc. in the lot. Take it to the HNL ('hole notha' level): Try using the raw model, with the raw to turbo Lora (then you aren't limited to 1024 pixels) along with 12 steps instead of 8, with this combination of sampler/scheduler which I never even tried until a few months ago: exp_heun_2_x0_sde linear_quadraticYes, exp heun 2 x0 sde - ooh it just rolls off the tongue. they need a marketing team. It seems to work great in Minimax H3 with beta scheduler. v3<span style="background-color:rgb(26, 27, 30);color:rgb(193, 194, 197);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">The best D has gotten even better with this finetuned version trained with a new high-resolution dataset that was targeting low noise stage for even better details. This was trained for another 6 hours of GPU time; total 11000 steps. The only other LoRA used with some of the showcase images is my <a target="_blank" rel="ugc" href="https://civarchive.com/models/2769362"><span style="background-color:rgb(26, 27, 30);color:rgb(34, 139, 230);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">Abercrombie Lora</a><span style="background-color:rgb(26, 27, 30);color:rgb(193, 194, 197);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">; no cum lora necessary but you can use one with it of course. No filter bypass node or lora needed.v2Trained for 10k steps total of 10 hours on my 5090, same dataset but focused training on the low noise to improve details.v1It only needed 4.5 hours to train v1 on my 5090 against the raw model and my 200 image dataset, therefore it can be used with both the turbo model (all of my showcase images are turbo, 8 steps) or the raw model. I didn't get a single backwards dickhead, and only a couple detached penises that were easily fixed by modifying the prompt. My 20-image showcase is only cherry picked to exclude the images that didn't show D in them. It is that good!It knows many A-list celebs already without images or loras, but their safety feature added more clothing when I tried to use my lora with them. You can use something like this <a target="_blank" rel="ugc" href="https://civarchive.com/models/2728234/krea2filterbypass">Krea2FilterBypass</a> if mine doesn't give you any results, but showcase 3 images were all created with no other loras except some used my Abercrombie style lora.

Penis by CoachBate Krea 2  (THE BEST D)

fine-tuned details for total of 10k steps

Files with SHA256: bf4e0ce4679f27791406eb5c68e4dcb325b8abc8cfb186533f4ddddc8ea9db6d