On October 13, 2026, Higgsfield AI announced that advances in OpenAI models have allowed the company to produce videos exceeding 10 minutes, a notable increase from an earlier practical limit near one minute. CEO and co‑founder Alex Mashrabov attributes the change to GPT‑6 Astra and related tools that help maintain visual and narrative consistency across longer sequences.
Longer videos with GPT‑6 Astra
Higgsfield’s creative agents now handle repetitive production tasks such as generating multiple shot variations and can call external tools to perform complex 3D modeling work, the company says. Alex Mashrabov says this tool-calling ability addresses a key challenge for creative AI: keeping scenes consistent over extended runtimes.
The company has applied the workflow to its own instructional channel, which Mashrabov says has almost half a million subscribers. He reports that producing those longer instructional videos previously required substantial manual effort, and that GPT‑6 Astra helped build a workflow capable of delivering videos over 10 minutes.
Images, code and a path to voice-to-video
For promotional content, Higgsfield highlights GPT‑Image‑2.5’s ability to render legible text and infographics within images, which the startup says helps small businesses create more precise visual materials without large production budgets. Mashrabov emphasizes that the model can reliably place text and data overlays on imagery.
On the software side, Higgsfield combines roughly 200 creatives and 100 engineers in product development. Mashrabov says Codex gives creatives “the power of a really strong software engineer” to turn ideas into prototypes. He adds that GPT‑6 Astra can supply production-ready code, reducing the friction between prototype and product.
Looking ahead, Higgsfield plans to explore a “voice to movie” workflow using GPT‑Live‑1, aiming for an experience where spoken direction is turned into visual output in near real time. Mashrabov describes OpenAI’s voice model as “best in class for voice understanding and just for regular communication.”
Higgsfield positions these capabilities as tools to help smaller creative teams tackle larger projects; the company cites concrete internal use cases and the scale of its development team as evidence of progress.
Original source: OpenAI News