OpenAI Has Finished the Sora 1 to Sora 2 Handoff in the U.S.
OpenAI says Sora 1 has been unavailable in the United States since March 13, 2026, with the service now opening directly into Sora 2 for U.S. users. That is more than a cosmetic rename. It ends the old split experience where some users could still move between the original Sora web product and the newer Sora 2 workflow.
The immediate practical consequence is simple:
- You can no longer switch back to Sora 1.
- Image generation inside Sora goes away with Sora 1.
- OpenAI wants still-image work to move into ChatGPT Images.
- Existing Sora 1 content is only exportable for a limited window before permanent deletion.
That makes this a product consolidation story, not just a feature update. OpenAI is narrowing Sora’s role around video creation while moving image generation into the broader ChatGPT surface.
What Sora 2 Looks Like Right Now
The newer Sora experience is materially more usable than the old “generate and start over” pattern. According to OpenAI’s current help documentation, Sora 2 now includes a built-in editor on web and iOS that lets users trim, stitch, reorder, extend, reprompt, and remix clips without leaving the app.
For creators, the biggest upgrades are:
- Storyboards on web for ChatGPT Pro users
- 25-second storyboard generations on the web for Pro
- 10- and 15-second standard generations in the main composer
- Clip stitching up to 60 seconds
- Integrated timeline editing instead of exporting every rough cut to a separate video editor
OpenAI also says Sora supports image creation on the web at sora.com, but that image workflow is now distinct from the old Sora 1 image-generation experience. In the sunset FAQ, the company explicitly says that once Sora 1 is removed, users should continue making images in ChatGPT instead.
Why This Matters for AI Video Creators
The old Sora product was easy to talk about because it looked like a single multimodal playground. The new setup is more specialized:
- Sora becomes the video-first product
- ChatGPT becomes the default place for still-image generation
- Editing moves closer to the generation step instead of living entirely in post-production tools
That is a sensible product split, even if it removes one convenience from Sora 1 users who liked having image generation inside the same interface.
For AI Photo Labs readers, the more important shift is workflow maturity. Sora 2 now looks less like a demo surface and more like a real iteration environment, especially for people building short cinematic sequences, stitched social edits, or rough previsualization reels. That also lines up with what we are seeing across the broader AI video market: tools are competing less on “can it generate a clip?” and more on how quickly you can iterate from idea to usable cut.
What Existing Sora 1 Users Should Do Now
If you still have older Sora 1 work, OpenAI’s guidance is to export your data as soon as possible. The company says Sora 1 generations, likes, and remixes will not remain available forever, and that users will get an email before the final export window closes.
The safest migration path is:
- Export legacy Sora data now.
- Rebuild any repeatable image workflows inside ChatGPT Images.
- Move ongoing video work into Sora 2.
- Re-test any old prompt templates that depended on Sora 1 behavior.
That last point matters. A sunset like this usually changes more than UI. It changes defaults, edit flow, duration choices, and sometimes what kinds of prompts or inputs feel native to the product.
Our take: OpenAI is trading feature sprawl for a cleaner product split. That is good for day-to-day usability, but it also means creators who used Sora 1 as an all-in-one image-plus-video sandbox need to rebuild part of their workflow. The upside is that Sora 2 now has enough editing and storyboard control to feel closer to a production tool than the original release ever did.