# Higgsfield takes AI video from 1 minute to over 10 > Source: > Published: 2026-10-14 12:00:00+00:00 [OpenAI](https://openai.com/index/higgsfield-longer-videos-voice) ### Higgsfield takes AI video from 1 minute to over 10 Which summary reads better? Pick one — models revealed after.Both summaries are AI-generated. Higgsfield now generates AI videos over 10 minutes long using GPT-6 Astra, moving beyond the prior 1-minute ceiling. For production teams, the meaningful shift is orchestration: longer generations make end-to-end video workflows more viable, but they also raise the bar for consistency, latency, review, and cost control across extended multimodal outputs. AI video generation just jumped from a one-minute ceiling to over ten minutes per clip, which collapses the stitching-and-continuity overhead that currently dominates any long-form video pipeline. If you're building video tools, you can now generate coherent multi-minute scenes in a single pass instead of orchestrating dozens of short clips, and the paired image-precision and live-voice models mean you should plan for multimodal, near-interactive filmmaking workflows rather than batch render-and-splice.