OpenAI says its Sora product is no longer available as of April 26, 2026. Sora was OpenAI’s video-generation product; its later Sora 2 release added generated audio and was presented as more steerable and physically realistic. This review explains what the releases offered and where OpenAI said they fell short. The capabilities below are OpenAI’s descriptions, not results from independent testing.
What was OpenAI Sora?
Sora was OpenAI’s generative video product. Its releases changed over time: the original Sora research overview focused on generating video from text and images, while the December 2024 Sora Turbo announcement described a more specific set of creation tools and output limits. OpenAI introduced Sora 2 in September 2025 as a video-and-audio generation model. Those release descriptions are historical, not specifications for a service available now.
In its Sora 2 announcement, OpenAI’s Sora Team wrote, “Video models are getting very good, very quickly.” That is the team’s broad characterization, not a measured performance statistic.
What could Sora do?
Text, images, and editing tools in earlier releases
OpenAI’s original research overview described creating video from text, animating still images, extending generated clips, and filling in missing frames. The December 2024 Sora Turbo announcement described remixing and blending generated material, along with a storyboard tool for shaping a video through a sequence of prompts or scenes. The exact capabilities varied by release; these descriptions should not be treated as a single, unchanged feature set.
#1 Best Overall
Sora Turbo’s historical output specifications
OpenAI said the December 2024 Sora Turbo release could generate videos up to 1080p resolution and 20 seconds long, with multiple aspect ratios. Those were specifications for that historical release, not independent quality findings or current service limits.
Sora 2’s audio and control claims
OpenAI described Sora 2 as able to generate synchronized dialogue and sound effects alongside video. The company also said the model offered improved steerability and more accurate physics than prior video-generation systems, with sharper realism and a broader range of styles. These are vendor-reported claims; the reviewed sources do not establish independent benchmarks comparing Sora with other systems.
What were Sora’s limitations?
OpenAI acknowledged shortcomings in its own descriptions of earlier Sora releases. Its December 2024 launch announcement said the deployed model could produce unrealistic physics and struggle with complex actions over longer durations. The original research overview also identified problems with cause and effect, spatial directions such as left and right, and precise instructions about events unfolding over time, including camera movement.
Those weaknesses matter when a video needs consistent physical interactions, a specific sequence of actions, or exact spatial and camera direction. OpenAI’s later descriptions of Sora 2 emphasized improvements, but the reviewed sources do not provide an independent test showing how much those issues were reduced or whether they were resolved.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
How did OpenAI address safety and provenance?
OpenAI described safeguards for generated video, including content moderation, restrictions on some uploads, and stringent protections for content involving minors. Its Sora safety materials also discussed controls around people’s likenesses and characters, harmful content, generated audio, and reporting or deletion. OpenAI described visible and invisible provenance signals, including C2PA metadata, intended to help identify generated content.
These are measures OpenAI said it implemented, not guarantees that every unsafe output, misuse, or provenance error would be prevented or detected. Provenance signals can help provide context about a file, but they do not by themselves establish that all video will carry complete or reliable verification.
Is Sora still available?
No. OpenAI’s Sora pages state that the Sora product is no longer available as of April 26, 2026. Earlier access and pricing references are historical: the September 2025 Sora 2 announcement described an initial free rollout, ChatGPT Pro access to Sora 2 Pro, and planned API availability, while the December 2024 Sora Turbo announcement described subscription and usage details for that release. Neither announcement establishes that those offers remain active.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How does Sora compare with other video generators?
A current, evidence-based comparison cannot be made from the available product descriptions alone. Comparing Sora with a currently available generator would require up-to-date primary-source information about access, pricing, output controls, audio, duration and resolution limits, safety and provenance, and commercial terms. OpenAI’s historical feature claims do not establish how Sora would rank against other products today.
Recommended Free Tools
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
