sora concurrency limit

[Rajan-007]

I’m limited to just 2 concurrent video-generation jobs using Azure Sora via AI Foundry. Additional attempts (third concurrent submission) fail outright even when I’ve hit capacity, degrading my build pipeline and UX.Please increase the concurrent generation limit for Sora . A bump to at least 10–15 concurrent jobs would align with typical Gen-AI workloads and dramatically improve throughput.

[amynic]

Hi @Rajan-007 Thank you for taking the time to share this information. Currently sourcing extra information on your ask in Azure AI Foundry.

Its worth noting that SORA usage via Open AI is still limited and not available via an API. In there FAQs they share that users get access to 2 concurrent video generations (and Pro users that goes to 5 max) https://help.openai.com/en/articles/9957612-generating-videos-on-sora#h_6655208237

Given these limits it may be worth considering how to manage concurrency on the app side possibly? - queue-based-processing, batching requests or retries etc

[Rajan-007]

hi @amynic , thanks for clarifying — but I think there’s a misunderstanding here. The 2–5 job concurrency cap you referenced applies to users running Sora directly in the ChatGPT product.

My request is specifically about Sora deployed through Azure AI Foundry, where it’s exposed as an API for integration into applications. In that environment, the hard limit of 2 concurrent jobs is blocking my production workload. A bump to at least 10–15 would bring it in line with typical GenAI service patterns and allow us to handle video generation at scale without queuing delays or failed jobs.

Could you confirm if there’s a different quota path for Azure AI Foundry deployments of Sora vs. the ChatGPT usage tier?