AI Models Create Music Videos with Budget Constraints and Tool Autonomy

Jul 16, 2026

AI Summary

Two AI models, Claude Fable 5 and GPT-5.6 Sol, were tasked with autonomously creating music videos for the song 'Uptown Funk' under specified budgets. Each model utilized various tools and techniques, resulting in different outputs and spending patterns, highlighting the capabilities and limitations of current AI technology in creative tasks.

Claude Fable 5 and GPT-5.6 Sol were given the same song, a budget, and tools to autonomously create music videos.
Each model was tested at two budget levels ($25 and $100) across four runs, producing valid full-length videos with the original song.
The models operated independently, with Claude Fable 5 spending $48.60 and GPT-5.6 Sol spending $36.57 at the $100 budget level.
GPT-5.6 Sol utilized a mix of video models and an image-to-video pipeline in one run, while Claude Fable 5 focused on text-to-video generation.
The total costs included both generation spend and token costs, with Claude Fable 5's token costs accounting for 30-40% of each run's total, while GPT-5.6 Sol's token costs remained lower.
The project is open source, allowing others to experiment with different songs and models, and it highlights the current limitations and potential for improvement in AI-generated creative content.

ai musicvideo generationgpt-5.6claude fableartificial intelligence

AI Models Create Music Videos with Budget Constraints and Tool Autonomy

Related Stories

AI Helps Saxboy Billy Rise to Global Fame in 2023

ChatTJB offers human responses to user queries without AI technology

Anthropic outlines marking system for AI-generated content under EU regulations

Chinese tech companies halt AI companion services, leaving users heartbroken