ChatGPT adds virtual try-on and Favorites in global shopping update built on Images 2.5
OpenAI launched two shopping features in ChatGPT globally on Thursday: a virtual try-on tool for clothing and accessories, and a Favorites function for saving products. OpenAI says both use its newly launched ChatGPT Images 2.5 model. Pricing and plan availability were not disclosed in the source reporting.
OpenAI launched virtual try-on and Favorites in ChatGPT globally on Thursday, adding two shopping features to the assistant. The company says both are powered by its newly launched ChatGPT Images 2.5 model.
What launched
Virtual try-on. Users upload a selfie or a full-body photo to see how a piece of clothing or an accessory might look on them. A new "try on" button appears in ChatGPT's shopping results. Users can also upload an image of an item, such as a screenshot from a retailer's website, and ask ChatGPT to show it on them.
Favorites. Users can save products they find to a Library in the app. Saved items sit alongside their try-on images, according to OpenAI.
Other shopping workflows
OpenAI also pointed to existing ways ChatGPT can assist with shopping:
- Describe a style, then ask ChatGPT to find the pieces needed to complete the look.
- Upload photos of a celebrity's outfit and ask ChatGPT to find the items worn that are available for purchase.
The second workflow puts ChatGPT in territory that Pinterest and Google have long held for fashion inspiration and discovery that converts to retail sales.
The model behind it
OpenAI claims ChatGPT Images 2.5 produces more natural lighting and richer textures, follows editing instructions more reliably, and reduces image generation latency. These are company claims. No benchmark scores, latency figures, or independent evaluations were provided in the announcement coverage.
What was not disclosed
- Which ChatGPT plans (free or paid) have access to try-on, and whether usage limits apply
- Pricing or API availability for Images 2.5
- How uploaded selfies and full-body photos are stored, retained, or used for training
- Which retailers or product catalogs are supported in try-on results
Context
OpenAI has already pulled back from one shopping idea. An instant checkout feature did not perform well, and the company pivoted away from it. Elsewhere in the category, agentic AI startup Instinct began pushing product recommendations to users. Some users felt the proactive suggestions were an overreach, closer to ads than help. Google launched its own virtual try-on last year, so OpenAI is entering a space where a major competitor already has a product.
What this means
The features are a consumer-facing showcase for Images 2.5's editing fidelity. Try-on tests whether a model can keep a person's likeness and body proportions consistent while rendering fabric, fit, and lighting. That is a harder test than text-to-image generation from scratch.
The strategy is also notable. After the checkout setback, OpenAI is focusing on discovery and visualization, which come before purchase, rather than on transactions. Favorites and the Library also give ChatGPT a persistent, return-to surface for shopping, which a single chat session does not provide.
The open questions are privacy and distribution. Body photos are more sensitive than typical prompts, and OpenAI has not detailed its handling of them in the coverage available. Google also has a year's head start on try-on, along with deep retailer integrations through its shopping products. Whether users choose ChatGPT first for fashion discovery remains unproven.
Related Articles
GitHub Copilot team customizations now generally available for organization teams
GitHub announced that team customizations for GitHub Copilot are generally available for organization teams. The feature lets teams share the Copilot plugins they rely on without asking every member to repeat the same setup.
OpenAI Reports Up to 14% PPA Gain on AMD Flagship GPU IP From Model-Rewritten RTL
OpenAI reported on November 18, 2026 that its models achieved up to a 14% improvement in power, performance and area (PPA) on AMD flagship GPU intellectual property, measured against heavily optimized baselines. Some of the circuit rewrites were included in a design sent for manufacturing.
OpenAI models optimized circuits used in Jalapeño silicon and improved RISC-V core designs
OpenAI reports that its models optimized circuits used in Jalapeño silicon and improved RISC-V core designs. The company presents the work as progress in AI-assisted chip design with GPT-6 Astra.
Gemini Notebook now builds interactive websites from your notes with a "make me a site" prompt
Google's Gemini Notebook, formerly known as NotebookLM, can now turn a notebook of research into an interactive website from a text description. TechRadar tested it on King Arthur research and got a finished site in about five minutes, though the reviewer says claims need checking against sources.
Comments
Loading...