Google has released Veo 3.1 Lite, a significantly cheaper and faster AI video generation model through its Gemini API. The launch comes days after OpenAI shut down its Sora project. The new model costs just $0.05 per second for 720p video, making high-volume applications more viable. Google also announced a price cut for its mid-tier Veo 3.1 Fast model effective April 7, marking a strategic shift to capture developers needing scalable video features.
Google has introduced a new AI video model, Veo 3.1 Lite, which is priced at less than half the cost of its mid-tier Veo 3.1 Fast option. The model supports Text-to-Video and Image-to-Video in landscape and portrait formats at 720p and 1080p resolutions.
To put pricing in perspective, Veo 3.1 previously cost around $0.40 per second, while Veo 3.1 Fast ran $0.15 per second. Lite brings the floor down to $0.05 per second for 720p video, with adjustable durations of 4, 6, or 8 seconds.
The release follows OpenAI shutting down its generative video project, Sora, last week. Sora was reportedly burning $15 million per day, and the company announced it was pivoting to world simulation research, a move that impacted a $1 billion deal with Disney.
Chinese competitors like Kuaishou‘s Kling AI and Tencent‘s open-source Hunyuan Video have been offering comparable generation at much cheaper prices. The Chinese market has been competing aggressively on economics rather than just quality for some time.
On the professional end, tools like Utopai‘s PAI offer long-form cinematic storytelling with consistent characters and detailed storyboards for a higher cost. PAI signals where serious creators are heading, seeking control and not just generation.
Veo 3.1 Lite is positioned as an infrastructure play for developers who need to ship video features at scale. The company stated the model family gives developers flexibility based on their needs, with the message to builders being to pick their tier.
