OpenAI officially opens API for Sora 2 – the most advanced video and audio generation model

Last week, OpenAI released Sora 2, its most advanced video and audio generation model to date. The Sora 2 and Sora 2 Pro versions are available on the Sora app for iOS, ChatGPT, and Sora.com.

At DevDay 2025, OpenAI continued to expand access by officially making Sora 2 available through a developer API.

Create immersive videos from just natural language or images

According to OpenAI, Sora 2 is capable of generating detailed videos with smooth motion and natural sound, using only text descriptions or input images.

The model also demonstrates deep understanding of 3D space, motion, and continuity between scenes, making video creation from text more realistic than ever.

The Sora 2 API provides five main endpoints , including:

1. Create video – Start creating a new video from the prompt (description), can use a reference image or remix ID.
2. Get video status – Check video rendering status and progress.
3. Download video – Download the MP4 file after the creation process is complete.
4. List videos – List available videos, pagination supported.
5. Delete videos – Delete videos from OpenAI's storage system.

Two versions: Sora 2 and Sora 2 Pro

To meet diverse usage needs, OpenAI offers two variants of the Sora 2 model:

1. sora-2: Standard edition, low cost and high speed, suitable for quick experimentation, building drafts or creating social media content.
2. sora-2-pro: Higher end version, takes longer to render but gives cinematic quality, ideal for commercial or short film production.

Quality, limitations and price

Currently, the Sora 2 API does not support cameo (real human face) or converting input video to Sora style. However, OpenAI said these are features in development.

With the open API, Sora 2 promises to usher in a new era of AI-powered content creation, allowing developers, filmmakers, marketers, and creators to automate video production workflows at a professional level.

Samuel Daniel

Update 08 October 2025