Qwen Reasoning Intensity Test: 'xhigh' Mode Generates Massive Thought Tokens

SarcasticBaka · reddit · 2026-08-15

User testing Qwen3.8-27B (UD-Q4KXL quantization) observed a drastic increase in thought tokens when switching reasoning effort from 'medium' to 'xhigh'. In prompts to generate HTML games, 'medium' produced only a few thousand tokens, while 'xhigh' generated at least 15k, peaking at 40k for a Pacman example. The user questions if this behavior is expected.

Original post →

More from Models

Models channel →