If you’ve generated at least one image in Sora, you surely noticed it. Every picture has the same atrocious color palette. Every generation looks like a screencap from the final season of Breaking Bad. The washed-out yellows, browns, and oranges are everywhere.
The color palette you can see above was extracted with Adobe Color from a typical Sora-generated image, and it perfectly illustrates the problem. If you’ve ever created even one image using Sora, you’ll immediately recognize this distinctive, bad-looking, overly saturated tint.
The warped hands are not the sign of AI-generated images anymore. Today, the biggest giveaway of an AI-generated image from Sora (or ChatGPT’s built-in image tool) is this color palette. Even the Sora feed looks painful to watch with all these endless orange-toned memes of Elon Musk and Michael Jackson.
Sora feed right now
The Current State of Sora Images
OpenAI hasn’t shared much about the underlying technology powering Sora, but visually, it’s clear the images suffer from similar errors as earlier versions like DALL-E. Beyond strange colors, Sora-generated images often are noticeably grainy and suffer from heavy censorship.
But the biggest, most noticeable issue remains the orange-brownish color tint (let’s be generous and call it a “pumpkin orange”). Is there anything you can do about it? Absolutely. Here are the best methods to help you deal with Sora’s persistent Mexican sun filter.
Prompt the Image for a More Natural Color
One of the simplest ways to combat Sora’s “pumpkin orange” tint is by adjusting your prompt. While results can still vary, adding specific color-related keywords can steer the image generation in at least a bit cleaner direction.
Include the following keywords to reduce the tint:
cool lighting, blue tones, neutral color palette, cold lightning, cinematic lighting with natural colors, overcast daylight, blue palette
Avoid words like “warm,” “sunset,” or “vintage,” which often reinforce the orange tint.
But the prompting can only get you so far and often doesn’t yield correct results as this color palette is produced because of the model architecture. Here’s what else you can do.
Fix the Colors Instantly with Irfanview
This is by far the quickest and easiest method – perfect for beginners and takes less than 30 seconds. IrfanView is a popular lightweight, free image viewer that also includes some surprisingly effective editing tools.
The original Sora picture
Auto-adjusted for colors with Irfanview
To automatically fix the orange tint and restore more natural colors:
- Open the Image in IrfanView
- Go to Image -> Auto Adjust Colors
- Save your Image
That’s it! The change is instant, and in most cases, it does a great job neutralizing the heavy orange-brown tones with just one click.
If you want more control over the color correction process, IrfanView also offers manual adjustments. Just go to Image -> Color Corrections to open the editing panel.
Color corrections panel in Irfanview
From there, you can fine-tune individual RGB color channels, tweak the saturation, adjust the gamma (midtones), and more. It’s a simple interface, but it gives you enough precision to dial in a more natural, balanced look—especially useful if the auto-adjust feature doesn’t quite get it right.
Use Photoshop for Advanced Color Correction
IrfanView is fast and offers excellent adjustment features that should suffice for most use cases. But if you’re looking for even greater precision and control over colors, you should use Adobe Photoshop.
Here’s a simple way to correct the orange tint:
- Open your Image in Photoshop.
- Go to image -> Adjustments -> Color Balance
- In the Color Balance panel, move the sliders under the Midtones section. Typically, you’ll want to:
- Reduce the Red slider slightly (drag toward Cyan).
- Reduce the Yellow slider (drag toward Blue).
Adjust these until the Mexican filter fades away and the image colors appear more natural.
Adjusting the color balance in Photoshop
Additionally you can also switch to the Shadows and Highlights sections if needed, making minor tweaks to refine your results further.
When you’re satisfied, apply the changes and then save your image.
With these simple steps, you can quickly restore balanced colors to your images. Photoshop offers plenty more features, but even basic adjustments like these can dramatically improve the look of AI-generated images.
Ditch Sora for Stable Diffusion
Sora is a closed-source generative model maintained by OpenAI (controlled by Microsoft). It not only restricts what you can create but also enforces a narrow visual style.
There is a strong content bias and ChatGPT will lecture what you can’t generate. It doesn’t ask for your opinion but can refuse even innocuous prompts. This makes closed-source models unsuitable for creative work. It’s censored, and no one should tell you how you should do your art. It has nothing to do with ethical concerns. It’s a pure corporate arrogance.
In contrast, Stable Diffusion is open-source, customizable, and completely free of unnecessary restrictions. Here’s why it’s a superior alternative:
Why Stable Diffusion is better
- No Filters or Censorship
Generate any content you want, including NSFW or artistic expressions that closed systems block. - Thousands of Models and Styles
From photorealism to anime, pixel art to surrealism – there’s a model or LoRA for every niche. - No Rate Limits, Unlimited Generations
Everything runs locally on your own GPU, so there are no queues, daily limits, and paywalls. - Extreme Customization
You can use ControlNet, LoRAs, inpainting, upscaling, and more. Sora (or Midjourney) doesn’t even come close to this level of control. - Private and Secure
Your creations stay on your machine – no cloud, no tracking, no forced moderation. - User-Friendly Interfaces
Tools like A1111 and ComfyUI make Stable Diffusion accessible even to beginners while still offering deep customization for advanced users.
If you’re serious about AI-generated art, Stable Diffusion is the way to go.
Stable Diffusion can produce fantastic results, and interfaces such as A1111 (above) make it easy to use
Want to get started with Stable Diffusion? Check out this blog or pick up my Stable Diffusion Handbook. It’s a great introduction to generative AI, and it covers everything you need to know to start generating AI graphics. It will let you understand generative AI, learn the tools, and produce stunning results for social media, design work, game development – or your personal projects. Now is the best time to learn the AI art!
