OpenAI experienced a partial outage on Tuesday morning that created some issues for users trying to access ChatGPT, Sora, and the API, the company said on its status page.
The company started investigating the issues late on Monday night, but the partial outage has persisted through Tuesday morning. Around 5:30 a.m. PT on Tuesday, OpenAI says it identified the issue and started working to fix it.
However, at roughly 8 a.m. PT, OpenAI said full recovery across its services may take “another few hours,” meaning folks logging onto work on the West Coast of the United States won’t be able to access ChatGPT this morning.
While ChatGPT outages typically last just a few hours, Tuesday’s partial outage is notably long in duration. The company said users may experience “elevated errors and latency” when using ChatGPT.
When TechCrunch tried to access GPT-4o in ChatGPT on Tuesday morning, the chatbot responded with an error reading “Too many concurrent requests.”
Tuesday’s partial outage comes amid a flurry of announcements from OpenAI. At Apple’s WWDC event on Monday, the iPhone maker announced deeper integrations with OpenAI’s models. Also on Monday, an OpenAI spokesperson confirmed to CNBC that the company has reached $10 billion in annualized recurring revenue. Earlier on Tuesday, OpenAI CEO Sam Altman announced an 80% price cut for developers trying to access its o3 AI reasoning models in its API.
OpenAI has previously run into issues scaling the usage of its AI models to hundreds of millions of users, but that’s exactly what the company needs to do to meet its grand ambitions. Altman has previously said the company’s “GPUs are melting” to keep up with the demand for ChatGPT, indicating that OpenAI’s computing resources are spread thin these days. It seems that demand for OpenAI’s models is only continuing to rise.