# Sora Text-to-Video Demo

On February 15, 2024, OpenAI previewed Sora, a text-to-video model generating high-definition clips from prompts, later releasing it publicly in December 2024 and discontinuing it in 2026.

Sora was a text-to-video model and social media app developed by [OpenAI](https://www.wikiprompt.org/wiki/openai). Using artificial intelligence, it generated short video clips based on text prompts and could extend existing videos. OpenAI first previewed Sora on February 15, 2024, releasing example clips that demonstrated its ability to create high-definition, realistic scenes from simple descriptions. The model was later made publicly available in December 2024, and a second generation with social features launched in September 2025. Sora was discontinued in 2026, with the app shutting down on April 26 and the API on September 25 of that year.

The technology behind Sora adapted the architecture of [DALL-E 3](https://www.wikiprompt.org/wiki/dall-e), OpenAI's text-to-image model. Sora used a diffusion transformer, a type of [deep learning](https://www.wikiprompt.org/wiki/deep-learning) model that generates videos by denoising 3D patches in a latent space. The model was trained on publicly available and licensed copyrighted videos, though OpenAI did not disclose the exact sources. Sora could generate videos up to one minute long, but it had limitations in simulating complex physics, understanding causality, and distinguishing left from right. OpenAI restricted prompts involving sexual, violent, hateful, or celebrity content, as well as material featuring existing intellectual property.

## Background

Before Sora, several other text-to-video models had been developed, including [Meta's Make-A-Video](https://www.wikiprompt.org/wiki/make-a-video-release), [Runway's Gen-2](https://www.wikiprompt.org/wiki/runway), and Google's Veo. OpenAI had also released [DALL-E 3](https://www.wikiprompt.org/wiki/dalle-1) in September 2023, which generated images from text prompts. Sora built on this progress, aiming to push the boundaries of generative video.

The team named the model after the Japanese word for 'sky', signifying its "limitless creative potential." This naming reflected OpenAI's ambition to create a tool that could generate virtually any scene described in text.

## Initial Release and Public Access

On February 15, 2024, OpenAI first previewed Sora by releasing multiple high-definition video clips it had created. Examples included an SUV driving down a mountain road, an animation of a "short fluffy monster" next to a candle, two people walking through Tokyo in the snow, and fake historical footage of the California gold rush. OpenAI stated that Sora could generate videos as long as one minute. The company also shared a technical report detailing the training methods, and CEO [Sam Altman](https://www.wikiprompt.org/wiki/sam-altman) posted tweets responding to user prompts with Sora-generated videos.

Prior to public release, OpenAI provided limited access to a "red team" of experts in misinformation and bias for adversarial testing. The company also shared Sora with a small group of creative professionals, including video makers and artists, to gather feedback on its usefulness in creative fields.

As of December 9, 2024, OpenAI gradually made Sora available to the public for [ChatGPT Pro](https://www.wikiprompt.org/wiki/chatgpt) and ChatGPT Plus users in the United States and Canada. In February 2025, OpenAI announced plans to integrate Sora into ChatGPT, allowing users to generate videos directly from the chatbot.

## Sora 2 and Social Features

Sora 2 was unveiled on September 30, 2025, alongside an iOS app, with an Android app following two months later. The new version integrated social media features, making the app overtly similar to TikTok in style and features, as noted by outlets like Wired. All videos generated by Sora 2 featured a visible, moving digital watermark to prevent misuse. However, on October 7, 2025, 404 Media reported that third-party programs capable of removing the watermark had become prevalent within a week of release.

Sora 2 also introduced a safety watermark to help viewers distinguish between real and fictional content. The app's social features allowed users to share and interact with generated videos, positioning it as a competitor to short-form video platforms.

## Discontinuation

On March 24, 2026, OpenAI announced on X that it was discontinuing Sora in both the mobile app and the API. The Sora app was shut down on April 26, 2026, and the API was discontinued on September 25, 2026, marking the end of the Sora AI brand. OpenAI's partnership with Disney, which included a licensing agreement allowing Disney characters in Sora, was also ending.

The decision prompted British technology news website The Register to label OpenAI a "product-killer," following in the footsteps of companies like Google, [Amazon Web Services](https://www.wikiprompt.org/wiki/amazon-web-services), [Broadcom](https://www.wikiprompt.org/wiki/broadcom), Cloud Software Group, and Netscape. OpenAI did not provide a specific reason in its shutdown notice, but reports linked the decision to computation shortages, cost pressures, and a shift toward core enterprise products.

Following its public launch, Sora's worldwide users peaked at around one million before declining to fewer than 500,000. The service cost an estimated $1 million per day to operate due to the computational demands of video generation.

## Legal and Copyright Issues

In November 2024, an API key for Sora access was leaked by a group of testers on [Hugging Face](https://www.wikiprompt.org/wiki/hugging-face), who posted a manifesto protesting that Sora was used for "art washing." OpenAI revoked all access three hours after the leak and stated that "hundreds of artists" had shaped development and that participation was voluntary.

At launch, Sora 2 allowed copyrighted content by default unless copyright holders actively opted out. On October 3, 2025, OpenAI stated that a future update would give copyright holders "more granular control," but did not specify whether existing content would be removed. On October 6, the chairman of the Motion Picture Association criticized OpenAI's approach to copyright with Sora 2.

On December 11, 2025, the Walt Disney Company announced a $1 billion investment in OpenAI as part of a three-year licensing deal, allowing users on Disney+ to generate more than 200 copyrighted characters on Sora 2, including those from Disney Animation, Pixar, Marvel Studios, and Star Wars.

## Capabilities and Limitations

Sora's technology was an adaptation of DALL-E 3's architecture. According to OpenAI, Sora was a diffusion transformer, a denoising latent diffusion model with a transformer as its denoiser. Videos were generated in latent space by denoising 3D patches, then transformed to standard space by a video decompressor. Recaptioning, using a video-to-text model to create detailed captions, augmented the training data.

OpenAI trained Sora on publicly available videos and licensed copyrighted videos, but did not reveal the number or exact sources. The model learned to create 3D graphics from its dataset alone, as stated by researcher Tim Brooks. Fellow researcher Bill Peebles noted that the model automatically learned certain visual concepts.

Despite its capabilities, Sora had acknowledged shortcomings, including limited ability to simulate complex physics, understand causality, and differentiate left from right. These limitations were typical of generative video models at the time, which struggled with long-term consistency and physical accuracy.

## Impact and Legacy

Sora's preview in February 2024 generated significant attention in the [generative AI](https://www.wikiprompt.org/wiki/generative-ai) field, showcasing the rapid progress in text-to-video technology. It influenced discussions about the potential and risks of AI-generated media, including concerns about misinformation and copyright. The model's discontinuation in 2026, however, highlighted the challenges of sustaining resource-intensive AI products, even for leading companies like OpenAI.

Sora's legacy includes its contribution to the development of video generation models and its role in shaping public awareness of AI's creative capabilities. Its shutdown also served as a case study in the commercial viability of such technologies, as costs and competition mounted.

## External links

- [Wikipedia: Sora (text-to-video model)](https://en.wikipedia.org/wiki/Sora_(text-to-video_model))

---
Source: https://www.wikiprompt.org/wiki/sora-text-to-video-demo
License: CC BY-SA 4.0 (https://creativecommons.org/licenses/by-sa/4.0/)
Last updated: 2026-11-06T12:15:52.986783+00:00