Sora 2 Pro
ai model13 mentions· velocity: stableSora 2 Pro, also marketed as Sora Image, is an OpenAI video generation model first publicly documented on February 26, 2026. It generates video clips up to 60 seconds in length at 1080p resolution from text prompts or reference images, with an optional 4K upscaling pass available at lower frame rates. The model supports variable aspect ratios and includes a “Director Mode” for shot-list-based sequencing, confirmed in OpenAI’s technical addendum. Pricing is set at $0.25 per second of generated video, with an enterprise tier offering dedicated GPU allocation. Sora 2 Pro matters now because it represents OpenAI’s first explicitly commercial-tier video model, moving beyond research previews with production-oriented features like priority rendering and commercial-use licensing. Its release directly escalates competition with Runway’s Gen-4 and Pika 2.1, both of which announced comparable 60-second generation limits within the same quarter, positioning OpenAI strategically amid an accelerating industry push toward commercially viable AI video tools.
Two-hop subgraph: this entity, every entity it directly relates to, and every entity those neighbors relate to. Drag a node, scroll to zoom, click to inspect — or click any neighbor and re-center the atlas there.