Google releases Veo 3.1, adds it to Flow video editor (techcrunch.com)

Google announced Veo 3.1, an incremental but meaningful upgrade to its video-generation model that brings improved audio output, finer-grained editing controls, and stronger image-to-video fidelity. Building on May’s Veo 3, the update generates more realistic clips with better adherence to textual and visual prompts, lets users insert objects that blend stylistically into a scene, and adds synchronized audio to existing edit primitives — reference-image driven character launches, “first-and-last-frame” clip generation, and video extension from recent frames. The model is being rolled out across Google’s Flow video editor, the Gemini App, and via Vertex and Gemini APIs, and Flow has already powered over 275 million videos since launch. For the AI/ML community the release tightens the gap between generative research and production tooling: more reliable prompt adherence and audio-visual consistency make automated content creation usable in faster iteration cycles and downstream workflows. API availability signals easier integration into apps and pipelines, while granular edit controls reduce manual post-production. It also raises practical considerations — compute/latency for multimodal generation, dataset and safety auditing for realistic edits, and content-moderation needs as object insertion and upcoming object-removal features lower barriers to seamless video manipulation.