| Sora | |
|---|---|
| Developer | OpenAI |
| Initial release | December 9, 2024 (11 months ago) (2024-12-09) |
| Stable release | Sora 2 (September 30, 2025; 57 days ago (2025-09-30)) |
| Platform | OpenAI |
| Type | Text-to-video model |
| Website | sora |
| Part ofa series on |
| Artificial intelligence (AI) |
|---|
_artificial_intelligence_icon.png){kind=icon} |
Glossary |
Sora is atext-to-video model andsocial media app developed byOpenAI. The modelgenerates short video clips based onprompts, and can also extend existing short videos. After OpenAI showed some examples in February,[1] the first generation of Sora was released publicly forChatGPT Plus and ChatGPT Pro users in the US and Canada in December 2024,[2][3] and the second generation, calledSora 2, was released to select users in the US and Canada at the end of September 2025.[4] It integrates social media features.[5][6]
By default, the generator uses copyrighted material in its videos (unlesscopyright holders actively opt-out of having their content included), and videos contain a visible, movingdigital watermark to preventmisuse. However, only a week after Sora 2's release, third-party programs became prevalent which could remove the watermark.
Several other text-to-video generating models had been created prior to Sora, includingMeta's Make-A-Video,Runway's Gen-2, andGoogle'sVeo 3, the last of which, as of February 2024,[update] is also still in its research phase.[7]OpenAI, the company behind Sora, had releasedDALL·E 3, the third of its DALL-Etext-to-image models, in September 2023.[8]
The team that developed Sora named it afterthe Japanese word for 'sky' to signify its "limitless creative potential".[9] On February 15, 2024, OpenAI first previewed Sora by releasing multiple clips ofhigh-definition videos that it created, including anSUV driving down a mountain road, an animation of a "short fluffy monster" next to a candle, two people walking throughTokyo in the snow, and fake historical footage of theCalifornia gold rush, and stated that it was able to generate videos up to one minute long.[7] The company then shared a technical report, which highlighted the methods used to train the model.[10][11] OpenAI CEOSam Altman also posted a series of tweets, responding toTwitter users' prompts with Sora-generated videos of the prompts.
As of December 9, 2024, OpenAI had gradually made Sora available to the public for ChatGPT Pro and ChatGPT Plus users in the US and Canada. Prior to this, the company had provided limited access to a small "red team", including experts inmisinformation and bias, to performadversarial testing on the model.[8] The company also shared Sora with a small group of creative professionals, including video makers and artists, to seek feedback on its usefulness in creative fields.[12] In February 2025, OpenAI announced plans to integrate Sora into ChatGPT by letting users generate Sora videos from the chatbot.[13]
Sora 2 was unveiled on September 30, 2025, with aniOS app at the same time.[4] All videos generated by the model feature a visible, movingwatermark to preventmisuse behaviors of the tool.[14] The previous version of Sora also used a safety watermark to let the audience distinguish between real and fictional content.[15] Only 7 days later, on October 7, 2025,404 Media reported that third-party programs which successfully removed the compulsory watermark from Sora 2 videos had become prevalent.[16]
By default, the generator uses copyrighted material in its videos unlesscopyright holders actively opt-out of having their content included.[17][18]
Some internet users and online content creators, such as Hank Green,[19] have called the mobile app "SlopTok", a reference to both the popular mobile appTikTok and the popular termAI slop.[20] Many outlets, such as WIRED Magazine, have noted that the Sora 2 app is overtly similar to TikTok in style and features.[2][21][22]
In November 2024, anAPI key for Sora access was leaked by a group of testers onHugging Face, who posted amanifesto stating that they were protesting that Sora was used for "art washing". OpenAI revoked all access three hours after the leak was made public, and stated that "hundreds of artists" have shaped the development, and that "participation is voluntary."[23][24][25]
On its launch, Sora 2 allowed copyrighted content by default, unless copyright holders were to contact OpenAI and opt-out of the generation of their content on the platform.[26] On October 3, 2025, OpenAI stated that a future update to Sora 2 would give copyright holders "more granular control" over the generation of copyrighted content, though OpenAI did not state if existing content would be removed.[27] On October 6, the chairman of theMPA criticized OpenAI's approach to copyright on Sora 2.[28]
The technology behind Sora is an adaptation of the technology behindDALL-E 3. According to OpenAI, Sora is a diffusion transformer[29] – adenoising latent diffusion model with oneTransformer as the denoiser. A video is generated in latent space by denoising 3D "patches", then transformed to standard space by a video decompressor. Re-captioning is used toaugment training data, by using a video-to-text model to create detailed captions on videos.[11]
OpenAI trained the model using publicly available videos as well as copyrighted videos licensed for the purpose, but did not reveal the number or the exact source of the videos.[9] Upon its release, OpenAI acknowledged some of Sora's shortcomings, including its struggling to simulate complex physics, to understandcausality, and to differentiate left from right.[30] One example shows a group of wolf pups seemingly multiplying and converging, creating a hard-to-follow scenario.[31] OpenAI also stated that, in adherence to the company's existing safety practices, Sora will restrict text prompts for sexual, violent, hateful, or celebrity imagery, as well as content featuring pre-existingintellectual property.[8]
Tim Brooks, a researcher on Sora, stated that the model figured out how to create3D graphics from its dataset alone, while Bill Peebles, also a Sora researcher, said that the model automatically created different video angles without being prompted.[7] According to OpenAI, Sora-generated videos are also tagged withC2PA metadata to indicate that they are AI processed.[9]
In 2024, Will Douglas Heaven of theMIT Technology Review called the demonstration videos "impressive", but noted that they must have been cherry-picked and may not be representative of Sora's typical output.[12] American academicOren Etzioni expressed concerns over the technology's ability to create onlinedisinformation for political campaigns.[9] ForWired,Steven Levy similarly wrote that it had the potential to become "a misinformation train wreck" and opined that its preview clips were "impressive" but "not perfect" and that it "show[ed] an emergent grasp of cinematic grammar" due to its unprompted shot changes. Levy added, "[i]t will be a very long time, if ever, before text-to-video threatens actual filmmaking."[7] Lisa Lacy ofCNET called its example videos "remarkably realistic – except perhaps when a human face appears close up or when sea creatures are swimming".[8]
FilmmakerTyler Perry announced he would be putting a planned $800 million expansion of hisAtlanta studio on hold, expressing concern about Sora's potential impact on the film industry.[32][33]
In October 2025,The New York Times remarked that the release of the Sora 2 app in September 2025 was "jaw-dropping (for better and worse)" though also remarked that the app was a "social network in disguise" and "the type of product that companies like Meta and X have sought to build: a way to bring A.I. to the masses that people can share." The article expressed concern regarding the product's potential impact on society and its potential use to promote misinformation, disinformation, and scams.[34]
OpenAI came under controversy over character generation after Sora 2 produced several videos that featured copyrighted characters. The company stated it would work with rights holders to block characters from Sora at their request, giving copyright holders more control.[35] In October 2025, Japan'sContent Overseas Distribution Association submitted a request to OpenAI demanding that it stop using the copyrighted content of its member companies, includingStudio Ghibli andSquare Enix.[36][37]
Various estates of celebrities have threatened legal action against OpenAI's Sora 2 app, due to deepfake videos being created of their likeness, including celebrities that have passed away.[38] Family members of the late comediansRobin Williams andGeorge Carlin also urged OpenAI to take action against "hurtful videos" and to restrict deepfakes of their loved ones.[39] OpenAI restricted users from making videos of the lateMartin Luther King Jr. and gave estates the ability to opt out of those they represent.[40]
| Products |
| .svg){kind=logo} | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| People |
| ||||||||||||
| Joint ventures | |||||||||||||
| Related | |||||||||||||
| Concepts |
| ||||||
|---|---|---|---|---|---|---|---|
| Applications | |||||||
| Implementations |
| ||||||
| People |
| ||||||
| Architectures |
| ||||||
