Qwen 3.8 27B hits 70 t/s with specific config parameters
dsdt · reddit · 2026-08-20
Shared a "perfect config" for running Qwen3.8-27B-UD-Q6K on dual RTX 5060 Ti (32GB), achieving 68-70 t/s generation speed and 80% draft acceptance rate via parameters like speculative decoding, flash-attn, and batch size tuning.
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