Veo 3.1 is coming soon, and Google’s clearly aiming it right at Sora 2 with longer video support (www.techradar.com)

Google’s Veo 3.1 is starting to appear on third‑party AI video platforms (Higgsfield, Imagine Art, Envato) and brings several production‑oriented upgrades: native 1080p output, improved character consistency across frames, multi‑prompt “multi‑shot” generation, cinematic presets (drone, pans, tracking, lighting/tone), and better handling of photos, illustrations and cartoons. Where Veo 3 stood out for tight lip‑sync and integrated voice/music/effects, 3.1 extends that foundation to longer clips—sites advertise “30 seconds+” and leaks claim up to one minute—and finer scene control via multiple prompts to stitch together distinct shots from a single source. For the AI/ML community and creators this matters because it reduces two longstanding pain points—short clip length and character/scene drift—while offering higher fidelity and director-style controls that make generated video more useful for storytelling and prototype production. Technically, multi‑prompt sequencing and robust framewise consistency imply stronger temporal modeling and identity preservation, and the cinematic presets signal a push toward turnkey production workflows rather than one‑off demos. The upgrade also positions Veo 3.1 as a direct rival to OpenAI’s Sora 2, accelerating competition around quality, length, and controllability of text‑to‑video models—and raising similar downstream questions about IP, consent, and attribution as these tools move closer to production use.