Users Report Qwen3.8-Flash-Next Hallucinating an Eight-Year-Old AppImage Install
memeka · reddit · 2026-09-04
Reddit user memeka reported serious hallucination issues with Qwen3.8-Flash-Next: when running on a Mac, the model tried to "install" an Ubuntu AppImage from 8 years ago, and while pulling tensor metadata from Hugging Face it mentioned a "OpenD" that appears to be completely made up. The author emphasized this was not a low-quantization version — the minimum was Q4 and the model was running at 5bpw quantization — ruling out quantization as the cause of the hallucinations.
More from Models
- Grok web gets a cleaner redesign with a refreshed UI — XFreeze · 2026-09-05
- eyebench author says no v4, moving on to harder benchmarks — adonis_singh · 2026-09-05
- Astra-max claims vastly better intelligence-per-token even at low reasoning — adonis_singh · 2026-09-05
- Astra-max hits 95% on eyebench-v3 at half the cost of Sol-max, tokens ~3.8x fewer — adonis_singh · 2026-09-05
- pass@1 dead even, but Fable 5.1 wins pass@k over GPT 6 Astra — zainhas · 2026-09-05
- GPT-6 Astra's cybersecurity classifier blocks a non-security bug-hunting task in Codex — Elctsuptb · 2026-09-05