Seeking LLM Fine-Tunes for Blender and 3D Modeling

A Reddit user initiated a community discussion on the LocalLLaMA subreddit to inquire about modern LLM fine-tunes designed for Blender, 3D modeling tasks, and generating assets like .glb or gltf files. The user specifically asked about models in the size range of Qwen variants that feature vision support and can handle descriptive specifications and reference images. Automating 3D asset generation and editing through local language models represents a major frontier for developers looking to scale video game creation efficiently. Finding specialized fine-tunes could bridge the gap between text-based AI assistants and complex spatial design workflows. The query highlights specific use cases including generating 3D models from markdown specs with reference images, editing existing licensed models, rigging, custom animations, and converting formats like FBX to glTF. The author noted that frontier models consume excessive token budgets, making local mid-sized models a more cost-effective preference if properly specialized.