In May 2024, Google introduced its groundbreaking AI video generation model, Veo, at the I/O developer conference. This innovative tool allows users to create over 60 seconds of high-definition (1080P) video from text or image inputs. What sets Veo apart is not just its ability to generate stunning visuals but also the customization options it offers—users can adjust lighting conditions, camera angles, and color tones to suit their creative vision.
The initial demonstrations showcased remarkable capabilities in rendering intricate details like animal fur and lifelike motion effects. As technology progressed, so did Veo; by May 2025, Google launched the third iteration—Veo 3—which integrated with the Gemini platform and expanded access across 159 global markets. Subscribers could now produce three videos daily using this advanced model.
Fast forward to October 2025 when Google unveiled Veo 3.1—a significant upgrade that enhanced narrative control and audio integration while boosting visual realism. These improvements positioned it as a formidable competitor against other leading models such as OpenAI's Sora.
Interestingly enough, competition has intensified within this space. In December of the same year, ShengShu Technology released Vidu Q2 in China—a direct rival aiming for supremacy alongside Sora and Veo’s latest versions. Vidu Q2 boasts features like multi-entity consistency which enables creators to merge multiple reference images into coherent narratives without losing individual characteristics.
This leap forward highlights an essential trend: AI is evolving beyond mere content creation into a collaborative partner for human storytellers. Luo Yihang from ShengShu emphasized that their goal isn’t about replacing creativity but enhancing it—turning imagination into vivid storytelling experiences.
As we look ahead at these developments in generative AI video models—from Google's pioneering efforts with Veo to emerging challengers like Vidu Q2—the landscape promises exciting possibilities for filmmakers and content creators alike.