llama.cpp: Qwen3.8-27B Reasoning Effort vs Budget Confusion Clarified
bonobomaster · reddit · 2026-08-16
Reddit user bonobomaster warns that the reasoning selector in llama-server's web UI is just a reasoning budget (hard cap) and has nothing to do with Qwen3.8-27B's native reasoning effort.
Selecting values only truncates reasoning, while reasoning effort truly affects depth and analytical skills. It can be set via --chat-template-kwargs "{\"reasoningeffort\":\"medium\"}" (older versions) or --reasoning-effort medium (newer). Options: low, medium, xhigh (default).
More from coding & agent
- Moving agent work from GPUs to CPUs for durable, predictable actions — cramforce · 2026-10-16
- Jev tutorial: 106x cheaper, 180ms yes/no gates replace PASS/FAIL string matching — max_paperclips · 2026-10-16
- Stop giving agents raw API keys: prepaid balances and per-credential caps — Cold_Hall_5384 · 2026-10-16
- Six-month prediction: hiring for generic agent harnesses built from scratch will cool — yi111 · 2026-10-16
- Omarchy ISO wouldn't boot on NVIDIA, so he had Claude patch it — indragie · 2026-10-16
- Dev uses Grok to clean up 3D printer firmware, catches it sneaking in nonsense code — yacineMTB · 2026-10-16