One of the most exciting recent developments is Sora, OpenAI’s advanced video generation model. Designed to deliver custom, visually compelling results, Sora responds to carefully crafted prompts and guidance.

You can access it at Sora.com. This is a completely new product experience from OpenAI.

While anyone can browse the Explore Sora feed, only paying customers on ChatGPT can actually use Sora to create their own videos. Sora is currently up and running for U.S. customers who have an OpenAI Plus or Pro subscription. Plus users get 50 generations each month. Pro subscribers get even more flexibility—unlimited generations in slow-queue mode and 500 regular generations per month.

We will walk you through the fundamentals of interacting with Sora—from composing effective prompts, to refining stylistic choices, to navigating its unique features—in order to harness its full creative potential and transform your ideas into vivid, dynamic videos.

How does Sora work?

Sora can generate realistic videos from text prompts, supporting resolutions up to 1080p and durations up to 20 seconds. Users can create content in various aspect ratios and utilize features like storyboarding and asset remixing.

How to Prompt Sora: A Quick Guide

You can create videos, without audio, easier than every with Sora. How you frame your prompt, the text you give it, makes a difference in the video that is created. Here’s how to get the best results:

1. Be Explicit for Specific Results

• If you have a clear idea in mind, describe every detail.

  • You can use a tool like Prompt Perfect to write better prompts faster.

• Mention the scene, lighting, characters, actions, colors, or even the mood.

• Example: “A beach scene that captures the essence of a bright, sunny day. The setting should feature clear blue water, golden sand, and a child joyfully building a sandcastle. Include details about the child's expressions and actions, the texture of the sand, and the playful presence of seagulls soaring in the sky. The overall tone should evoke feelings of happiness and the carefree spirit of summer. Structure the output in a descriptive paragraph that immerses the reader in the scene, emphasizing sensory details such as sights, sounds, and feelings.”

2. Be Broad for Creative Freedom

• If you’re open to interpretation, keep your prompt general.

• Sora will fill in the gaps with its own creative decisions.

• Example: “A magical forest at dusk” or “A futuristic cityscape.”

3. Experiment with Styles

• Add style cues like “cinematic,” “cartoonish,” or “minimalist” to guide Sora’s output.

• Example: “A cinematic view of a sunset over rolling hills.”

4. Iterate and Refine

• If the first result isn’t perfect, adjust your prompt.

• Fine-tune details or add clarifications based on what you see.

Features of Sora

The model released today is called Sora Turbo. It’s a new high-end accelerated version of the original Sora model. It has all the capabilities of the research model with generating videos from text, animating images, and video-to-video features like remixing into new styles, extending forwards and backwards in time, and more.

It’s important to note this is a new model that will make mistakes. It is the first step in tooling creative people with capabilities they couldn’t access from a computer before.

Sora Explore Feed

This is a feed of videos shared by the community to see outputs from Sora. Learn about techniques, methods, and just great videos from other users.

Sora Library

This is where you can see all the videos you’ve generated and bookmarks from the explore feed.

Sora Composer

Describe a video with text or upload an image. Play with features like aspect ratio, resolution options, length duration, variations of the same video, and presets for aesthetics and styles in a reusable component like stop motion and balloon world.

Sora Storyboard

This is an advanced creative tool built into Sora. You can direct a video with actions across a sequence in a video editing timeline.