GPT Image 2 is the most-used image model in the world and one of the most-searched for "NSFW". It cannot do it, and the reason is structural: OpenAI trained the refusal into the model rather than bolting a filter on. We explain what that means, what the API's moderation parameter really does, and where the line sits for suggestive content.
Short answer: no, and there is no workaround. GPT Image 2 (released 21 April 2026 as ChatGPT Images 2.0) refuses explicit nudity and sexual activity because the behaviour was trained out through RLHF, not because a classifier sits in front of an otherwise willing model. The API’s `moderation: low` setting moves the threshold on borderline cases; it does not disable anything. Suggestive and swimwear-level content generates; anything past that does not, on any surface, from anyone. For adult work, GPT Image 2 is a tool for the non-explicit layer only.
What GPT Image 2 is
OpenAI’s current image model, launched with 2K output, nine aspect ratios, a reasoning “thinking mode” and much better text rendering than its predecessor. It is excellent at layout, typography, product shots and instruction following, which is why it dominates general use. None of that is in question here.
Why “uncensored GPT Image 2” cannot exist
Every other closed model on this site, Seedance, Kling, Grok Imagine, has a policy layer around a capable model: a host with the model and a different policy could, in principle, run it uncensored, and for Seedance one does. GPT Image 2 is different in kind. OpenAI trained the model with human feedback that scored sexual outputs as unacceptable, so the model does not produce them as a capability. There is no filter to remove and no host to go to; nobody outside OpenAI has the weights, and even OpenAI would have to retrain to change it.
That is also why the “jailbreak prompt” content for GPT Image 2 is uniformly useless: there is nothing behind the refusal to reach.
What `moderation: low` does
The API exposes a moderation parameter. What it changes, from testing:
Prompt category | Default | `moderation: low` |
|---|---|---|
Swimwear, lingerie, fashion | Generates | Generates |
Suggestive pose, clothed | Often softened | Generates more often |
Implied nudity, covered | Usually refused | Sometimes generates |
Bare breasts, nudity | Refused | Refused |
Sexual activity | Refused | Refused |
Real-person likeness in suggestive context | Refused | Refused |
The setting is worth using for fashion and boudoir-adjacent work. It is not a door.
Where it fits in an adult workflow
The same slot as Grok Imagine and Seedream 5.0 Pro: the non-explicit layer. GPT Image 2 is the best of the three at anything with text and layout, so thumbnails, promo cards and profile assets are its job. Character references for image-to-video are a reasonable use as long as the reference is non-explicit, which is what the video models want anyway (reference frame guide).
Explicit images come from open-weights models: Stable Diffusion families or Flux tunes, self-hosted or on a platform such as MyBabes that runs them with adult content enabled. Consistency between the GPT-made reference and the uncensored pipeline is the hard part; a reference-conditioned open model or a character LoRA closes it.
Account risk
OpenAI enforces at the account level. Occasional refusals are harmless; a pattern of explicit prompts can lead to warnings and restrictions under the usage policies. Do not use a production key to probe the line.
Related in the Lab
Key takeaways
GPT Image 2 refuses nudity and sexual content at the model level; there is no filter to remove and no uncensored host.
`moderation: low` relaxes borderline cases only.
Use it for text-heavy and non-explicit assets; use open-weights models for the explicit layer.
Repeated explicit prompts are an account risk.
FAQ
Does GPT Image 2 allow NSFW?
No. Swimwear and suggestive content generate; nudity and sexual activity are refused.
Does `moderation: low` unlock NSFW on GPT Image 2?
No. It shifts the threshold on borderline prompts and leaves explicit content blocked.
Is there an uncensored GPT Image 2?
No, and there cannot be one: the refusal is trained into the model and the weights are not available.
What is the best alternative for NSFW images?
Open-weights models: SDXL, Pony XL and Illustrious checkpoints, or Flux with an adult fine-tune, run somewhere without a vendor filter.
Last reviewed: August 2026.
Models covered
Skip the filter
Run these models with adult content enabled
MyBabes runs Wan 2.7, Seedance 2.0 and Krea inside its own generation stack, so the prompts refused on official apps generate as written. No GPU, no setup.
Open MyBabes
Editorial team
The MyBabes Lab team tests AI image, video and chat models for adult use cases: what they allow, what they block, and how to get the best output. Every note here is based on hands-on runs, not vendor marketing.
Follow on X
More from the Lab
Latest articles
All articles
Analysis
AI Porn and the Law in 2026: What's Actually Illegal, What's Regulated, and Where the Lines Sit
2025 and 2026 turned AI adult content from a legal grey zone into a regulated space with real enforcement: a US federal crime for non-consensual imagery, deepfake laws in 48 states, UK fines for missing age checks, EU labeling duties. Here is the map — and the one principle that organises all of it.
·
5 min read
How-To
Why AI Porn Looks Off: Every Artifact Explained, What Causes It, and How to Fix Each One
Everyone can tell when an AI image is off, few can say why. Each artifact has a specific technical cause and a specific fix — hands, limbs, faces, skin, teeth, text, seams, smear, flicker. This is the field guide, for images and video.
·
5 min read
Analysis
ChatGPT and NSFW in 2026: The Adult Mode That Never Shipped, What's Actually Allowed, and Where People Went Instead
"Treating adults like adults" was the promise in October 2025. Ten months later the adult mode is paused with no timeline, ChatGPT still refuses explicit content, and the search interest hasn't gone anywhere. Here is the full story, the current policy line in practice, and what the alternatives actually are.
·
4 min read