---
cited_by:
  - '[[thoughts/inference/deepseek-v4.1-serving#4-pr-reading-order|dsv41 §4]]'
date: '2026-11-11'
kind: pr
local: []
merged: '2026-09-10'
number: 56208
source:
  - https://github.com/vllm-project/vllm/blob/b47b01cf3383d50a5fae8569cadcbd0736076234/vllm/tokenizers/deepseek_v41_encoding.py
  - https://github.com/vllm-project/vllm/blob/b47b01cf3383d50a5fae8569cadcbd0736076234/vllm/models/deepseek_v4_1/common/mm_preprocess.py
  - https://github.com/vllm-project/vllm/blob/b47b01cf3383d50a5fae8569cadcbd0736076234/rust/src/chat/src/renderer/deepseek.rs
  - https://github.com/vllm-project/vllm/blob/b47b01cf3383d50a5fae8569cadcbd0736076234/rust/src/chat/src/renderer/deepseek_v41/tests.rs
  - https://github.com/vllm-project/vllm/blob/b47b01cf3383d50a5fae8569cadcbd0736076234/rust/src/chat/src/multimodal/expand.rs
  - https://github.com/vllm-project/vllm/blob/b47b01cf3383d50a5fae8569cadcbd0736076234/vllm/tokenizers/deepseek_v41.py
  - https://github.com/vllm-project/vllm/blob/b47b01cf3383d50a5fae8569cadcbd0736076234/rust/src/parser/src/tool/deepseek_dsml/deepseek_v41/tests.rs
  - https://github.com/vllm-project/vllm/blob/b47b01cf3383d50a5fae8569cadcbd0736076234/rust/src/parser/src/tool/deepseek_dsml/deepseek_v41/structural_tag.rs
state: merged
tags:
  - vllm
  - vllm-pr
  - new-architectures
title: '#56208 [Model][Frontend] Support DeepSeek-V4.1-Flash in Rust and Python frontends'
topic: new-architectures
url: https://github.com/vllm-project/vllm/pull/56208
created: '2026-11-11'
modified: '2026-11-11'
published: '2026-11-11'
pageLayout: default
slug: hinterland/papers/maps/prs/56208
permalink: https://aarnphm.xyz/hinterland/papers/maps/prs/56208.md
generator:
  quartz: v4.6.0
  hostedProvider: Cloudflare
  baseUrl: aarnphm.xyz
full: https://aarnphm.xyz/llms-full.txt
---
Extract the Rust and Python frontend portions of [#56201](https://github.com/aarnphm/aarnphm.github.io/issues/56201) onto main for separate review: tokenization/rendering, reasoning and tool parsing, tool grammar, image preprocessing, and config adapters. Pins `llm-multimodal` to `24c676d`; enables… · full dump: [[thoughts/inference/vllm-prs/56208|#56208]]

