---
cited_by:
  - '[[thoughts/inference/qwen3.8-runtime#2-delta-vs-predecessors|qwen4next §2]]'
  - '[[thoughts/inference/qwen3.8-serving#3-kv-cache--state-what-lives-where-per-request|qwen4next §3]]'
  - '[[thoughts/inference/qwen3.8-serving#4-pr-reading-order-vllm-main-git-show-sha|qwen4next §4]]'
date: '2026-11-11'
kind: pr
local:
  - file:///Users/aarnphm/workspace/vllm/vllm/config/vllm.py
merged: '2026-09-07'
number: 55272
source:
  - https://github.com/vllm-project/vllm/blob/d9105ea8001e0a6d77a96327d17515bb5791fb36/vllm/models/qwen4_exp/nvidia/qsa.py
  - https://github.com/vllm-project/vllm/blob/d9105ea8001e0a6d77a96327d17515bb5791fb36/vllm/models/qwen4_exp/nvidia/ple_layer.py
  - https://github.com/vllm-project/vllm/blob/d9105ea8001e0a6d77a96327d17515bb5791fb36/vllm/models/qwen4_exp/nvidia/mtp.py
  - https://github.com/vllm-project/vllm/blob/d9105ea8001e0a6d77a96327d17515bb5791fb36/vllm/models/qwen4_exp/nvidia/model_state.py
  - https://github.com/vllm-project/vllm/blob/d9105ea8001e0a6d77a96327d17515bb5791fb36/vllm/models/qwen4_exp/nvidia/indexer_qsa.py
  - https://github.com/vllm-project/vllm/blob/d9105ea8001e0a6d77a96327d17515bb5791fb36/vllm/models/qwen4_exp/nvidia/model.py
  - https://github.com/vllm-project/vllm/blob/d9105ea8001e0a6d77a96327d17515bb5791fb36/vllm/config/vllm.py
  - https://github.com/vllm-project/vllm/blob/d9105ea8001e0a6d77a96327d17515bb5791fb36/vllm/model_executor/layers/mamba/gdn/qwen_gdn_linear_attn.py
state: merged
tags:
  - vllm
  - vllm-pr
  - new-architectures
title: '#55272 [Qwen3.8-Flash-Next] Remove torch.compile for NVIDIA implementation'
topic: new-architectures
url: https://github.com/vllm-project/vllm/pull/55272
created: '2026-11-11'
modified: '2026-11-11'
published: '2026-11-11'
pageLayout: default
slug: hinterland/papers/maps/prs/55272
permalink: https://aarnphm.xyz/hinterland/papers/maps/prs/55272.md
generator:
  quartz: v4.6.0
  hostedProvider: Cloudflare
  baseUrl: aarnphm.xyz
full: https://aarnphm.xyz/llms-full.txt
---
Fixes [#54688](https://github.com/aarnphm/aarnphm.github.io/issues/54688) · full dump: [[thoughts/inference/vllm-prs/55272|#55272]]

