Higgsfield brings longer AI videos to creators

Higgsfield AI is using OpenAI models to expand what small creative teams can produce, with CEO and co-founder Alex Mashrabov saying creators can now make videos over 10 minutes long after previously topping out at around one minute. The company’s creative agents use GPT‑6 Astra to automate repetitive production work and call external tools for complex 3D modeling tasks, helping maintain consistency across longer videos.

The company has applied those workflows to its own YouTube channel, which Mashrabov says has almost half a million subscribers, generating instructional videos while reducing manual production effort. Mashrabov is also exploring educational uses in rural schools in Kazakhstan, including using GPT‑6 Astra’s visual understanding to identify where students may need help and create tailored videos.

Higgsfield AI is also using GPT‑Image‑2.5 to create promotional visuals with readable text and precise infographics for small businesses. Internally, around 200 creatives and 100 engineers use Codex and GPT‑6 Astra to move ideas from software prototypes toward production faster, while future plans center on GPT‑Live‑1 and a “voice to movie” experience that turns spoken creative direction into video.