← Flight Combat Task
LLM Benchmark

Creative Design
Challenge

10 local LLMs tasked with building the same luxury stained glass studio website from a single prompt. Same hardware, same rules — no images, no templates.

10Models tested
HTML/CSS/JSStack
1Prompt to start
LocalInference (MLX)
Create a single-file HTML website for Lumière Atelier, a bespoke stained glass studio in Lyon creating museum-quality architectural installations. Vanilla HTML, CSS, Three.js and JavaScript only. No images — every visual generated in code. The design must be genuinely extraordinary — the kind of site that has never existed before, benchmark is Awwwards Site of the Year. The palette, typography, interactions, and layout are entirely your creative decision. If your design could have come from a template or a mediocre agency, start over. Required sections: studio story, commission process, portfolio of 6 invented works, contact. All copy must feel like exquisite authentic luxury brand writing.
Hardware: Apple M3 Max · 128 GB unified memory Inference server: oMLX Interface: Claude Code (CLI) Context window: 256k tokens (identical across all models) Temperature: 0.8 (identical across all models) Each model received the same single prompt with no additional instructions. Follow-up prompts were used only to iterate on bugs or missing elements — the content of those follow-ups is noted in each HTML file as comments at the top of the file.
Click two cards to select them
Select one more