Qwen3.8-Flash-Next local quants hallucinate 'corrupted context' errors, reports Reddit user
arkham00 · reddit · 2026-09-03
A Reddit user reports that Qwen3.8-Flash-Next GGUF quantizations (Q4KM, Q3KXL, IQ4XS, with/without MTP, short and long contexts) run locally via llama.cpp on a Mac M2 Max 96GB frequently hallucinate "garbled text" — the model declares tool instructions or .md files corrupted, then runs extensive git/system checks even though the files are fine. The issue seems to be in the model's own context, with occasional stray Chinese characters and misspellings.
Notably, the model is self-aware about its errors and tries to recover; in one case it claimed a tool was corrupted, but when the user told it to just run the tool, it worked — the model apologized and continued. The author finds the base model strong on instruction-following, much better than predecessors, despite the preview-quality quirks.
More from Models
- Open models below 4-bit struggle at self-iterating 3D asset generation, HF dev finds — mervenoyann · 2026-10-21
- Researcher Finds Two Windows 0-Days, First in 76 Minutes with a Frontier Model — moyix · 2026-10-21
- Dev finds Gemini Argon's debug-log analysis beats Fable 5.1 — ShikharMurty · 2026-10-21
- ChatGPT Gets Stuck Screaming "AAHH" in an Infinite Loop After a Simple Prompt — AutomaticScene · 2026-10-21
- Open-weights Ornith-1.5-122B debuts with 84.5% SWE-bench Verified, runs on a single DGX card — _akhaliq · 2026-10-21
- qev-0.8B: a tiny local decision model that picks moves in 3.5ms on 1.3GB — fajfas3 · 2026-10-21