Qwen3.6 35B - TXT vs Markdown vs HTML vs HTML+CSS

In three linesComparative test of Qwen 3.6 35B across output formats (raw text, markdown, HTML, HTML+CSS). Markdown achieves best quality (78/100 per ChatGPT-4o) with 1,496 output tokens in 23s. HTML+CSS generates 10,290 tokens in 82s but lower quality score (58/100). Measurements include reasoning tokens, throughput, and total time.

Summary generated by Claude — human-verified