Higgsfield AI Video Generation Hits 10+ Minutes with GPT-6 Astra

A significant leap in AI video duration has been announced, moving the frontier from one-minute clips to productions over ten minutes long. This shift, powered by new multimodal models, fundamentally changes the scope of stories and concepts creators can generate directly from prompts, opening a path toward voice-controlled filmmaking.

What did Higgsfield announce?

According to an announcement from OpenAI News, the Higgsfield platform has integrated new AI models that dramatically extend video generation capabilities. The core advancement is the use of GPT-6 Astra for long-form video creation, pushing the maximum length of AI-generated videos past the ten-minute mark. Alongside this, Higgsfield has incorporated GPT-Image-2.5 for more precise visual control and introduced GPT-Live, described as a foundational step toward voice-driven filmmaking. This suite of tools represents a concerted push toward more complex, director-led AI media production.

Why does longer AI video generation matter?

Extending AI video from roughly one minute to over ten minutes is not merely a quantitative change but a qualitative one. It transitions the technology from a tool for creating short clips, demos, or social media snippets into a viable medium for narrative shorts, educational explainers, detailed product showcases, and segments of longer-form content. This expansion allows for proper scene development, character arcs, and more sophisticated pacing, which are essential for storytelling and professional communication. It moves AI video generation closer to being a practical tool for filmmakers, educators, and marketers.

What is the "path to voice filmmaking" with GPT-Live?

The announcement mentions GPT-Live as establishing a "path to voice filmmaking." While specific technical details are not provided in the source, this suggests a development trajectory where a creator could use spoken commands—rather than typed prompts or complex editing interfaces—to direct an AI film. In practice, this could mean verbally instructing the AI on scene changes, character actions, camera angles, or editing choices, making the creation process more intuitive and fluid. It points to a future where the barrier between a creator's vision and the generated media becomes thinner and more conversational.

How can creators explore long-form AI video in Grafy?

For creators interested in pushing the boundaries of AI-generated narrative and documentary content, Grafy's Film Studio app is the direct workspace to explore. It is designed to plan, sequence, and assemble AI-generated scenes into cohesive short films and documentaries, making it an ideal environment to experiment with the kind of multi-scene, longer-format storytelling that Higgsfield's ten-minute videos enable. Users can structure a narrative, generate or import visual and audio assets, and compile them on a timeline.

Furthermore, the entire Grafy ecosystem supports the iterative, multimodal workflow that complex video projects demand. You can prototype visuals in Creator, storyboard ideas in Graph, source research with Reader, and even score original music in Music Studio for your productions. For technical deep dives into the models enabling these advances, Grafy Chat provides a side-by-side interface to query the latest research and reasoning models about video generation techniques.

This announcement follows a pattern of rapid capability expansion in the multimodal AI space, similar to recent leaps in reasoning models like Gemini 4 Argon and specialized agent efficiencies like GPT-6.1 Sol. For creators, the practical implication is that the tools for AI-assisted filmmaking are evolving from novelty to genuine utility.

Frequently asked questions

How long can Higgsfield AI videos be now?

According to the source, Higgsfield's integration with GPT-6 Astra allows it to generate AI videos that are over ten minutes long. This is a significant increase from a previous limit of approximately one minute, enabling much more substantial narrative and explanatory content.

What is GPT-Live for filmmaking?

GPT-Live is described as establishing a "path to voice filmmaking." It suggests a future development where creators might use voice commands to direct AI-generated films, making the process more interactive and intuitive compared to traditional text prompting or manual editing.

Can I make long AI videos in Grafy today?

Grafy's Film Studio app is built for assembling multi-scene AI films and documentaries. While the specific ten-minute Higgsfield model isn't integrated, the app's structure is designed for the long-form, narrative workflow that these new capabilities enable, allowing you to combine scenes, narration, and music into a complete film.

---

Source: OpenAI News — Higgsfield takes AI video from 1 minute to over 10