Open AI today Announced the release of Surah 2Its latest video generation model, which also includes audio from the now -made audio.
It is made with a new start of a new IOS app is easily called SurahWhich allows users to enter and edit with AI videos with themselves and their friends "Chemo" Feature – Although Openi says there are strong reservations and identification safety measures to prevent one’s identity from entering AI videos without their consent or approval.
In addition, the company said that a SORA application is in programming interface (API) works that will allow third -party developers to pipe the new SORA 2 model into their video editing applications, unlock new, more excellent and professional editing capabilities, and generally allow the Frontier Video Generation Model to advance the new direction.
In addition, Open says the Android mobile app is also working. Wired Magazine (Which my wife operates) First it was reported that the Surah 2 AI mobile app was being developed a day before the Open AI official announcement.
Chat GPT users in the United States and Canada will first raise their hands on the model and app, though the Open has said it intends to extend access to other countries in the coming days and weeks. Surah 2 is available free with the limits used for all users, while Chat GPT Pro Subscribers Get access to high quality "Surah 2 Pro" Model; Chat GPT Plus users do not get more additional benefits than free levels.
The release was presented during a direct series on YouTube, hosted by Open AISORA team researchers Bill Pables, Rohan Sahai, and Thomas Demson, who run the audience through a roadmap for model capabilities, chemo features, and upcoming tools.
One step forward in video AI
It is difficult to remember it right now, but Openi shouted the world with his realistic AI video when he teased his original Surah video model in the early 2024, just to rotate the role slowly in a small number of creative partners until he released it in December 2024.
Until then, the whole place of AI video generation had gone largely, and in the subsequent months, with Startups Runway, Luma, Killing, Higgsfield and many other photoralistic AI video models of many other rivals, with many better quality and baked in audio generation, with a lot of better quality and baked audio generation.
The Open AI described the original Surah as its “GPT -1 Moment” for the video, a point when the video generation for the first time began to show signs of refreshment.
On the contrary, Surah 2 has been made closer to the “GPT -3.5 Moment”, which is marked by more modern physics, realism and control.
Model can handle complex measures such as gymnastics routines or pedal board tricks, while obeying physical principles such as speed and happiness.
Unlike the previous system, which can “teleport” basketball in a hoop, when the shot is missed, Surah 2 is starting to return realistic health. It also compathers dialogue, background audio and sound effects, which create video audo experiences compatible in style from photovirialstick to mobile phones.
Is a standout feature "Arrival"By which users allow themselves or friends to enter the scenes created by their own or friends after one -time recording to achieve imitation and sound.
During the series directly, the use of chemo is fully optimized, which is protected from certification challenges to prevent imitation, and is renewable at any time. Here is an example of a chemo Surah 2 video that features an open AI researcher, which is discussed with Big Foot.
Surah app
The new Surah app serves as the main entry point for the model. It enables consumers to make videos and remarks, browse personal feed and socially support. Users can leave themselves through chemos in others videos, remix trending creations with their turns, and guidance and head through indicators.
The app is based on the invitation at the time of launch, the aim of the opening is to ensure that people are involved with friends. According to the company, the feed is designed different from ordinary social media platforms.
Instead of spending maximum time in scrolling, the app prefers the discovery of videos of the possibility of affecting creation. The content is weighed by people that the user follows or interacted with it, and the personal language can be adjusted by natural language instructions.
Surah is available for free on iOS, which is subject to the scope of computing uses. Over time, the Open plans are planned to offer an optional salary level to produce additional videos when demand is high.
Chat GPT Pro users will access the high quality “Surah 2 Pro” model through Sura.com and, finally, in the app. An Android version is in development.
Identification Protection and Camos
The chemo system is also central to identifying identification on the platform.
Setup and verification: Users recording a short video and dynamic audio sample in the app. Open AI systems confirm samples with audio challenges to ensure authenticity and prevent imitation.
Permissions control: Once confirmed, a user chooses who can use his chemo in breeds: only self, select contact, mutual or each. These permissions can be edited at any time.
Is needed: Users can adjust how the model offers them, adding prices such as clothing or tone deception, or adding playful stylish variations.
Rights to cancel and delete: At any time, a chemo owner can revoke access. They also have the right to delete any video in which they have an example. Openi describes that consumers can control the property style on their identity within the system.
Safety for adolescents and ordinary audiences
Openi emphasizes the characteristics of young safety and welfare at the time of launch.
Anti Domscroling Design: For users under the age of 18, the app disables unlimited script through default. Instead, the feed naturally stops after a certain number of videos, which requires coalodown before resuming. Even adult users are bent if they are trapped in the extended passive scroll, as the app prefers more creativity than consumption.
Protection of content for minors: When a rated chemo recording or image detects a potential minor in the image, the tough threshold is applied. This ensures that subsequent generations are filtered against additional types of harmful or inappropriate output.
Privacy default: Teen accounts come up with strong privacy settings, how their example can be used, limiting discovery by adults, and adding obstacles to unrelated contact.
Parents’ control: Parents can use chat -linked tools associated with Chat GPT to accommodate teen experience. These control feed limits can overwhelm the algorithmic personalities, manage chemo permits, and can directly ban messaging.
These steps reflect the intention of balanceing the experiences with the welfare of the anneity, acknowledging concerns about drug addiction and harmful social dynamics.
Safety and provision
Along with these identification concerns, the Open Eye System Card has outline wider security arrangements:
Input and output moderate using a multi -modal rating.
Restrictions on creating public personalities or photovirialstick similarities without consent.
Automatic detection of harmful materials applies to the social feed of the Surah with additional testing.
Provision features such as C2PA metad dataVisible moving water marks on downloaded videos, and verification of internal tracing AI-generation content.
The company contributed with the external Red Team’s testers to pressure the system against the category such as extremism, Iranian, self -harm and political manipulation.
Roadmap: Story Board and API
Beyond the app, Openi highlighted new features in development for SORA.com, including storyboard tools that will allow creators to control how the shot opens through a video shot. According to the current series, the capabilities of the story board are expected within weeks.
The company also confirmed that an API will be ready for Surah 2 in the coming weeks, which will open the model developers who want to integrate video generation into their tools and editors.
“There is a long tail for use matters where people do amazing things, where we may not want to make better grain editing controls, but others can be,” the presentation explained.
Altman’s reflection
Openi CEO Sam Altman Seat His point of view in a personal blog post With the launch, he described Surah as a “chatter for the moment of creativity” and said that the initial testers found the way to force the chemo feature to contact others.
The Altman also confessed to the possible downward. He wrote, “We are aware of how addictive such a service is, and we can imagine that it can be used for bullying.” He warned against the risk of “RL-OPTMized Sloped Feed”, but emphasized that the team has worked to design the results.
He presented numerous guidelines: Improving long -term user satisfaction, improving users, controlling their feeds, preferring creation, and helping people achieve their goals. If consumers do not feel that their lives have been improved after several months, the Altman said that the open would make significant changes, or completely shut down the service.
What comes next?
For now, Surah 2 is available through the app and SORA.com, with API access soon. Surah 1 turbo is active, and users’ past creations will be accessible in their libraries.
Open Frame Surah 2 2 not only as a tool of entertainment and creativity but also as a step towards the world’s imitation and wider ambitions in the AI system that can interact with physical reality. As the Surah team said in its announcement, the system is still incomplete, but it indicates progress towards “reality imitation”, while offers a new way to experience users with video.
A combination of new tools, a social app designed to prefer creation, a strong identification protection framework, and an API that aimed to developers views that Openi views both as a user product and a platform. Even though it will not only depend on its technical capabilities, one of the important places of digital creativity will also depend on how effectively the company manages safety, fitness and consumer trust in the next months.