ChatGPT Images 2.6 Adds Faster AI Image Generation, GIF Creation, and Saved Photo References
OpenAI has rolled out Images 2.6, a major update to ChatGPT’s image tools that makes AI image creation faster, more flexible, and more personal. The update brings quicker image generation, improved editing consistency, animated GIF support, and a new way to save personal photos as reusable references in future chats.
According to OpenAI, Images 2.6 is available worldwide across desktop, mobile, and web. The feature is included on every ChatGPT plan, including the free tier, and is also available through OpenAI’s API and Codex.
One of the biggest improvements is speed. OpenAI says image generation is now 50% faster than before, which should make the experience feel smoother for users creating artwork, product mockups, profile images, social media graphics, or visual ideas on the fly.
The update also improves image editing consistency. That means ChatGPT should do a better job preserving important details when users request changes. For example, if you edit the background of a portrait, the person’s face, clothing, and overall appearance should remain more stable. The same applies to product images, where details like shape, color, branding style, or design elements are less likely to change unexpectedly during edits.
A standout addition in Images 2.6 is the ability to create animated GIFs directly in ChatGPT. Users can describe the animation they want, or upload a still image and ask ChatGPT to bring it to life. For example, someone could request a waving cartoon character, a blinking reaction image, a moving logo concept, or a simple looping animation for social media.
OpenAI says users can adjust playback speed after generating a GIF, then save or share the result. If changes are needed, users can simply ask for them in the chat. However, ChatGPT currently cannot edit an animated GIF that a user uploads. It can create new GIFs from prompts or still images, but uploaded animated GIF editing is not supported.
The GIF feature is still rolling out, so it may not be visible in every account immediately. OpenAI has also not yet published clear limits for GIF generation across different ChatGPT plans.
The second major feature is saved photo references. This allows users to upload a personal image, give it a name, and reuse it later by mentioning that name in a future prompt. OpenAI’s example involves saving a dog photo as “Pepper.” Later, the user can ask ChatGPT to create a watercolor portrait of @Pepper without uploading the same photo again.
This could be useful for people who frequently create images of the same pet, character, product, outfit, or personal item. Instead of uploading the image repeatedly, users can save it once and bring it back into future conversations with a simple mention.
ChatGPT may also offer to save a photo when a user describes or works with it. Once saved, the image remains connected to the account for later use.
However, this new feature also raises privacy questions. OpenAI’s current help information does not clearly explain where saved photo references are stored, how users can view them, or how they can delete them. That is an important detail, especially for anyone uploading personal photos, pictures of family members, or images of children.
Generated images appear to follow a clearer rule: they are saved under the Images section and are removed when the conversation that created them is deleted. But for saved reference photos, OpenAI has not yet provided the same level of detail about management and removal.
The ability to save real-world photos and combine them with AI-generated animation also creates potential misuse concerns. A saved image of a real person could theoretically be used to place them in scenes they never appeared in, especially when paired with GIF creation. OpenAI has not yet given detailed public guidance on how it prevents abuse involving animated images of real people.
OpenAI says images made with ChatGPT include provenance signals, including Content Credentials and an invisible SynthID watermark. These tools are designed to help identify AI-generated content. Still, GIFs have not been explained separately in the same level of detail, so users may want to be cautious when creating or sharing animated content involving real people.
Anyone using the new saved photo feature should think carefully before uploading images of others. Photos of children, friends, relatives, coworkers, or private individuals should only be saved or reused with proper consent. Since uploads are part of ChatGPT conversations, users should also review their data settings to control whether their content may be used to improve AI models.
Images 2.6 makes ChatGPT more powerful for everyday creative tasks. Faster generation, more reliable edits, GIF creation, and reusable photo references could make it easier to design social posts, personalize artwork, create fun animations, or build consistent visual concepts over time.
At the same time, the update highlights the growing importance of privacy controls, consent, and transparency in AI image tools. The new features are convenient, but users should understand what they upload, what gets saved, and how personal images may be reused before making saved references part of their regular workflow.
