MikhbarMIKHBAR
Artificial Intelligence

Higgsfield Boosts AI Video Creation Beyond Ten Minutes

Higgsfield AI has expanded its AI video generation capabilities from around one minute to over ten minutes using advanced technology.

Higgsfield Boosts AI Video Creation Beyond Ten Minutes

Expanding Video Lengths Beyond One Minute

Higgsfield AI is scaling professional video and image creation to make it accessible to anyone with an idea. According to [OpenAI News](https://openai.com/index/higgsfield-longer-videos-voice/), previous video generation limits topped out at around one minute, but users can now create videos over ten minutes long.

CEO and co-founder Alex Mashrabov explains that longer videos enable smaller creative teams to tackle more ambitious projects. With models like GPT-6 Astra and GPT-Image-2.5, a single creator can execute high-end fashion campaigns from start to finish without needing massive production infrastructure.

invideo Option A art card
Image related to the report from OpenAI News · Source: OpenAI News

Maintaining Consistency with Creative Agents

The creative agents employed by Higgsfield AI handle repetitive production tasks, such as generating multiple variations of a specific shot. With the integration of GPT-6 Astra, these agents are capable of calling external tools to manage complex 3D modeling tasks.

Mashrabov notes that this capability directly addresses one of creative AI's most difficult hurdles: maintaining visual consistency across longer video sequences. Evidence of this workflow is visible on the official [Higgsfield AI ⁠](https://higgsfield.ai/) YouTube channel, which features instructional content generated with the help of the new models.

Enhancing Visuals for Small Businesses

For small businesses producing promotional material, clear text and precise infographics are often as crucial as the underlying imagery. GPT-Image-2.5 facilitates this kind of detailed content creation for businesses operating without large production budgets.

Mashrabov highlights that GPT-Image-2.5 is a state-of-the-art model capable of reliably rendering text and infographics directly on top of imagery, marking a significant advancement for automated promotional material generation.

Playco Astra customer story art card - Option B
Image related to the report from OpenAI News · Source: OpenAI News

Bridging Prototyping and Production

The development of the Higgsfield AI product relies on collaboration between a team of approximately 200 creatives and 100 engineers. Codex provides the creative team with advanced software engineering capabilities, allowing them to turn creative concepts into functional software prototypes independently.

Previously, moving prototypes into production required significant intervention from front-end and back-end engineers. Mashrabov reports that GPT-6 Astra reduces this friction, allowing the company to develop and ship refined products much faster.

Higgsfield > Card Image
Image related to the report from OpenAI News · Source: OpenAI News

Looking Toward Voice-to-Video Filmmaking

Looking ahead, Higgsfield AI aims to introduce voice-to-video filmmaking, enabling creators to describe concepts and watch them materialize in real time. Mashrabov refers to this long-term vision as voice-to-movie, where AI interprets spoken instructions to render visual narratives.

By combining advancements in planning, tool calling, image generation, and voice comprehension, the company hopes to make film creation feel as intuitive as describing an idea aloud, with AI handling the complex execution steps in between.

Sources

  • OpenAI NewsHiggsfield takes AI video from 1 minute to over 10

Continue chronologically

Related entity coverage