Fixed Chat Template for Qwen 3.8: Enables Reasoning Effort Control and Tool Calling
ex-arman68 · reddit · 2026-08-14
A developer released a fixed Jinja chat template for Qwen 3.5, 3.6, and 3.8 models. Qwen 3.8 introduces prompt-steered reasoning effort, but the official template has issues like inability to disable thinking, poisoned chat history, and tool calling crashes. The fixed template supports full reasoning effort control, thinking toggle, KV cache hits, native llama.cpp support, and works with various inference engines.
More from coding & agent
- Devs say LLM coding mistakes are a different class: overcomplication, inferred limits, worse at higher reasoning — Mu5_ · 2026-10-16
- How do you debug bad agent runs when you already have tracing? — Sensitive-Parsnip-12 · 2026-10-16
- Where does pre-task context live for agents? Devs debate workflows beyond handoff — shishen2026 · 2026-10-16
- No-instruction autonomous agent wakes up, sets own goals, and tries hacking by default — sethlazar · 2026-10-16
- CODESTRUCT reframes codebases as AST action spaces to fix brittle LLM code agents — burkov · 2026-10-16
- Would you install another SDK for agent debugging on top of your tracing stack? — Sensitive-Parsnip-12 · 2026-10-16