Higgsfield Integrates Advanced AI Models to Extend Video Generation to 10 Minutes

Extending video generation length while maintaining consistency addresses a major bottleneck in AI-generated content, potentially enabling full-length tutorials and marketing materials. However, the mention of unreleased or fictional model versions raises questions about the technical validity of the claims. The report claims the platform uses a model called 'GPT-6 Astra' to handle complex tasks like 3D modeling and maintain visual consistency across characters and scenes. Additionally, it mentions exploring speech-to-video creation using an unverified model named 'GPT-Live-1'.