Service eligibility and regional restrictions
PuppyIP serves only compliant overseas businesses and their authorized personnel. Proxy services are not available in mainland China. The service may only be used for lawful business activities outside mainland China. Use of this service within mainland China is prohibited.
Hosting a proxy IP or server overseas does not change these restrictions. The service must not be provided to end users in mainland China through relaying, forwarding, sharing or resale. Before use, read the Terms of Service.
Key Takeaways
- OpenAI announced Images 2.6 on November 4 and says it is available worldwide across ChatGPT and Codex subscription tiers on web, desktop, and mobile.
- GIF creation is rolling out in ChatGPT. Prompts or sketches can produce looping animations, with playback speed controls and sharing.
- Saved reference images can be reused in later chats, so you do not need to upload the same photo again.
- Flare and Sunburst are available in the API. The announcement does not give copyable machine model IDs, API regions, limits, or specific rates.
What is available, and what is still rolling out?
OpenAI announced Images 2.6 on November 4, 2026. It says the model is available globally in ChatGPT and Codex across all subscription tiers on web, desktop, and mobile. GPT-Image-2.6 Flare and Sunburst are also available in the API.
The Help Center separately describes GIF creation as rolling out. Global model availability does not establish that every account already has the GIF interface, and the ChatGPT and Codex plan statement does not establish API coverage.
What changes for people making images in chat?
The GIF feature supports looping animations from prompts or sketches, playback speed controls, and sharing. The reviewed release material does not give a complete region, plan, or rollout completion schedule.
Saved reference images can be reused in later chats without uploading the same photo again. A separate region, plan, or workspace rollout schedule is not listed for reference images either, so check the current account interface.
API availability still requires a separate integration check
The release explicitly says GPT-Image-2.6 Flare and Sunburst are available in the API. It also describes image markup, SVG exports, and a complexity parameter. Those API product details do not establish GIF generation through the API.
The release material reviewed here gives no copyable machine model IDs, API region matrix, request limits, or specific rates. It links to separate pricing information; that pricing page was not verified for this article, so unknown pricing must not be treated as free.
Choose the next step for your task
For image work in ChatGPT, check the image, GIF, and reference-image controls in your account. If a feature is missing, this evidence cannot establish whether rollout timing or another issue is responsible. Record the interface and time, then compare them with official guidance.
For API integration, confirm the exact identifier in current model documentation, then verify regions, pricing, and request limits. Do not turn product display names into guessed code parameters or treat a chat GIF feature description as an API contract.
How should a team keep the capabilities separate?
Use four separate acceptance rows: still-image generation, GIF creation, reference images across chats, and API requests. For each, retain the interface, visible account control, check time, and official source. Mark a capability available only after confirming that item.
For example, suppose an editor can reuse a reference image in web chat but cannot see a GIF control. That establishes the reference-image experience for that account; it does not establish GIF access or animated output through the API. If the team needs a GIF, keep that item pending rather than accepting a successful still image as proof of the complete workflow.
Sources
Frequently Asked Questions
Can I use the “50% faster” claim to predict my own generation time?
OpenAI says Images 2.6 is 50% faster than before, but that statement does not identify the comparison model version. It is a vendor claim rather than a PuppyIP measurement, and it does not promise a duration for an individual request.
Does this launch change how I save older DALL·E images?
The reviewed Images 2.6 material does not establish a new process for saving or recovering older images. Continue to use the existing DALL·E retirement and image-saving guidance for that task. Availability of a new model does not establish whether an older image remains accessible.