Benchmarking Qwen3.8 27B quantizations: 4-bit holds up, 1-bit collapses
stared
12 points
2 comments
August 26, 2026
Related Discussions
Found 5 related stories in 45.2ms across 4,560 title embeddings via pgvector HNSW
- Qwen3.8 27B scores 52 on Artificial Analysis anana_ · 329 pts · August 17, 2026 · 66% similar
- Qwen3.8 27B at 256K: 50 TPS on a 24 GB GPU pich · 37 pts · August 17, 2026 · 66% similar
- Qwen 3.8 27B is excellent, but it defaults to overthinking things bilsbie · 258 pts · August 16, 2026 · 64% similar
- Qwen3.8 27B kristjansson · 12 pts · August 14, 2026 · 56% similar
- Qwen 3.8 27B erdaltoprak · 1016 pts · August 14, 2026 · 56% similar
Discussion Highlights (2 comments)
kamranjon
Hmm, makes the 2 bit quants actually seem pretty reasonable…
Gerard22Aug
Yeah, there is latest and greatest trend: 0-bit quant. Absolutely outstanding performance, can produce as many tokens per second as you terminal may sustain with zero load on unified memory / GPU. Real winner!