OpenAI has unveiled its newest partnership with Higgsfield, an AI video generation platform. Following the integration of GPT-6 Astra, Higgsfield has successfully boosted the maximum length of AI-generated videos from a mere 1 minute to a substantial 10 minutes. Capitalizing on GPT-6 Astra's capacity to empower creative agents in invoking external tools to tackle intricate tasks, it guarantees uniformity in characters, settings, and visual elements across lengthy videos. Higgsfield has already implemented this relevant process on its YouTube channel, which boasts nearly 500,000 subscribers. This move has enabled the generation of instructional videos exceeding 10 minutes, thereby cutting down on manual labor. Moreover, Higgsfield has incorporated GPT-Image-2.5 to consistently render legible text and infographics, aiding small businesses in producing promotional materials at a minimal cost. Concurrently, leveraging OpenAI's Codex, Higgsfield has expedited product development, allowing creatives to directly transform ideas into software prototypes with diminished engineer participation. In the upcoming phase, Higgsfield is set to delve into real-time voice-to-video creation modes grounded in GPT-Live-1.