Deterministic prefill crash at a fixed token position on qwen4exp (Qwen3.8-Flash-Next) under multi-GPU layer split — 4 runtimes, FA on/off,

该问题出现在 llama.cpp / LM Studio 的 CUDA runtime 对 qwen4exp(Qwen3.8-Flash-Next)模型做长上下文 prefill 时,multi-GPU layer split( --split-mode layer )下会在固定 token 位置确





![[Bug]: Responses API streaming regenerates output item id/call_id in response.completed (breaks strict clients)](https://www.chat-gpts.plus/wp-content/uploads/2026/10/59834-4a72fa90-768x403.jpg)


