Sora 2
OpenAI
试用面板在模型正式上线后开放。
即将上线
该模型尚未开放，订阅邮件通知以第一时间收到上线消息。
规格参数
定价
TBD — pricing not yet announced
Coming soon
Sora 2 is OpenAI's second-generation text-to-video model and the most anticipated AI video release in the industry. The original Sora stunned the world in February 2024 with its ability to generate minute-long videos with complex camera movements and physically plausible scenes. Sora 2 builds on that foundation with substantially longer generation windows, 4K native resolution, and a new architecture designed to maintain narrative coherence across scenes — capabilities that push AI video into professional film and television territory.
能力说明
Sora 2 is expected to deliver several category-defining capabilities: 60-second video generation that can sustain a coherent narrative arc, native 4K output suitable for broadcast and large-format display, and enhanced world simulation that accurately models the physical behaviour of materials, crowds, and natural phenomena. Early reports also suggest improved character consistency across cuts — a feature critical for storytelling applications where the same character must appear in multiple scenes without drifting in appearance.
最佳实践
When Sora 2 launches, the most effective prompts are expected to follow a screenplay structure rather than a static image description. Instead of "a busy city street", write "open on a wide shot of a rain-soaked city street at dusk; a lone taxi pulls up and a woman in a red coat steps out, looking up at the neon signs". Temporal connectives — "then", "as", "while", "cutting to" — help the model understand narrative progression across the clip duration.
常见问题
How is Sora 2 expected to improve on the original Sora?
Sora 2 is anticipated to address the key limitations of the original Sora: longer generation windows (up to 60 seconds vs the original 20 seconds), native 4K output, improved temporal consistency for complex multi-character scenes, and stronger adherence to narrative prompts that describe events unfolding over time. OpenAI has also signalled improvements in physics accuracy for fluid, cloth, and particle simulations.
When will Sora 2 be available on Genace?
Sora 2 is not yet publicly available via API. Genace is on the early access list and will integrate it as soon as OpenAI opens the API to third-party platforms. Sign up on the Genace waitlist to receive an instant notification at launch.
Will Sora 2 support commercial use?
Based on OpenAI's existing Sora commercial policy, videos generated via the API are expected to be available for commercial use subject to OpenAI's usage policies. Specific licensing terms for Sora 2 will be confirmed once the API launches.
相关模型
Veo 3
Generate stunning 1080p videos with Veo 3 by Google DeepMind. The first commercially available Veo model features native audio generation, physics-accurate simulation, and up to 8 seconds of cinematic video from text prompts.
Veo 4
Veo 4 by Google DeepMind is the next generation text-to-video model featuring unprecedented realism, up to 4K resolution, and 60-second video generation. Join the waitlist to be first to access it on Genace.
Luma Ray 3
Luma AI
Luma Ray 3 brings 3D scene understanding and photorealistic ray tracing to AI video, producing cinematic camera motion and natural light behaviour that no other text-to-video model can match. Join the Genace waitlist for early access.