Veo 3.1 Ingredients to Video: More consistency, creativity and control

Source: DeepMind Date: 2026-01-13 URL: https://deepmind.google/blog/veo-3-1-ingredients-to-video-more-consistency-creativity-and-control/

Summary

Google updated Veo 3.1 with improved character and scene consistency when generating video from reference images, native 9:16 vertical output for mobile, and upscaling to 1080p/4K. The update emphasizes narrative coherence — consistent character identity across scenes and richer dialogue. No benchmark numbers; the post is a product capability showcase. Available across Gemini app, YouTube Shorts, Flow, Google Vids, and Vertex AI API.

Implications

The video-generation distribution thread. Veo’s real story is ecosystem breadth: one model powering YouTube Shorts, Gemini app, the Flow filmmaking tool, and enterprise Vertex AI. That’s the same distribution leverage Google uses for Search and Android — spread the model across properties rather than winning on a single benchmark. OpenAI’s Sora and Runway don’t have equivalent native distribution.

Portrait-mode video is the mobile wedge. Native 9:16 output is a direct push into the short-form content creator workflow. If Veo 3.1 can replace or augment CapCut-style tools for YouTube Shorts creators, Google closes the loop between generation and distribution in a way no competitor can match.

No benchmarks = product marketing. The lack of quantitative claims signals this is capability communication, not research. The competitive pressure is from Sora, Runway Gen-3, and Kling — watch for third-party quality comparisons on those against 3.1.

Watch:

• Creator adoption of Veo in YouTube Shorts versus Sora/Runway in the same period
• Whether Flow gains traction as a professional filmmaking tool or stays a demo
• Vertex AI enterprise video generation pricing and use-case fit