Users Complain Qwen 3.8 Outputs Are Dense and Hard to Read
parepeg · reddit · 2026-08-31
A user complains that Qwen 3.8 (27B and Flash Next) outputs are difficult to read, citing examples like using set intersection symbols instead of plain English and odd word choices like 'persona'. The author attributes this to an effort to minimize 'tokens per intelligence', which sacrifices human readability, and expresses concern that Qwen is following the same path as Claude 5 in becoming less readable.
More from Models
- OpenAI stealth-testing mysterious internal model 'Aeon' in Codex desktop app and Dots — cedric_chee · 2026-10-16
- Argon edges out astra as daily driver, but Claude still wins heavy coding work — iruletheworldmo · 2026-10-16
- NVIDIA ships FP4-quantized Qwen3.8-Flash for local use, trending on Hugging Face — nvidia · 2026-10-16
- RoboVista benchmark: Gemini-4 Argon hits 81.9% on robotics understanding — berkeley_ai · 2026-10-16
- Dev reports decision model jev beats gpt-6-luna on quality in real training pipeline — genmon · 2026-10-16
- Google won't publish Argon 4 limits in AI Pro—users compare experiences — Consistent-Text-2662 · 2026-10-16