Unsloth's Qwen 3.8 Q4KXL Fails in Long Context, Bartowski Holds Up
Healthy-Nebula-3603 · reddit · 2026-08-18
Comparative testing of Unsloth (Q4KXL) and Bartowski (Q4KM) Qwen 3.8 27B GGUF models revealed a critical failure in the Unsloth version. Beyond 60k tokens and after context compaction at 72k, the Unsloth model collapsed with broken logic and empty reasoning, while the Bartowski model remained stable.
More from Infra
- Large data centers employ only ~50-150 permanent on-site staff, Microsoft says — brezina · 2026-10-16
- Prediction: Opus 5.5-level models on laptops by 2027, backed by Qwen 27B data — cephaloform · 2026-10-16
- Oracle down 12% in a week, nearly 50% off June high as OpenAI exposure spooks market — alvelda · 2026-10-16
- US data center moratoriums top 400 — an AI bull explains why they're actually good for AI — kevinsxu · 2026-10-16
- Microsoft Research unveils Magnetar, the fastest zkVM at 10-39x speedups and post-quantum security — jedisct1 · 2026-10-16
- Bonsai 2 2B hits 120 tok/s on iPhone 18 Pro, outpaces Qwen 3.5 2B at half the memory — cephaloform · 2026-10-16