OpenAI puts Sora into ChatGPT
OpenAI announced plans to integrate its Sora video-generation model directly into ChatGPT so users can generate videos alongside text and images — but Sora 2 is unstable and faces compute, cost, and legal limits, so rollout looks constrained right now reported.
The Information’s reporting — echoed by outlets including Engadget — says OpenAI will keep Sora available as a standalone app even as its core generation capabilities are folded into ChatGPT. (engadget.com) App-store analytics cited in recent coverage put Sora’s standalone installs down roughly 45% in January 2026, a decline reporters link to early usage limits and engagement drops. (blockonomi.com) Independent trackers and community posts document Sora 2 instability: OpenAI retired Sora 1 on March 13, 2026, while users report jobs stuck at 99% completion, frequent “under heavy load” failures, and new accounts with video generation features disabled. (help.apiyi.com) OpenAI has baked provenance and safety controls into Sora 2 outputs, including C2PA metadata watermarks and explicit blocks on photorealistic depictions of public figures and other restricted content categories. (resotune.com) Microsoft has already wired Sora 2 into Bing Video Creator, demonstrating a licensing/distribution pathway outside OpenAI’s own apps and showing how partners can offload some deployment complexity. (neowin.net) Product analyses and changelogs list hard usage controls for Sora 2: daily generation quotas, video-length caps, and credit-based consumption tied to subscription tiers as published limits. (merlio.app) Industry reporting notes ChatGPT now handles roughly 900 million weekly active users and about 50 million paying subscribers, a scale factor analysts say could “spike costs and usage” for integrated video generation backends if rollouts are not throttled. (techcrunch.com)
