---
cited_by:
  - '[[thoughts/inference/kimi-k3-prs#4-pr-reading-order|kimik3 §4]]'
date: '2026-11-11'
kind: pr
local: []
merged: '2026-07-28'
number: 50104
source:
  - https://github.com/vllm-project/vllm/blob/2899dca8432d40632987b0ec24253a8fe6df2710/rust/src/parser/src/unified/kimi_k3.rs
  - https://github.com/vllm-project/vllm/blob/2899dca8432d40632987b0ec24253a8fe6df2710/rust/src/chat/src/renderer/kimi_k3/encoding.rs
  - https://github.com/vllm-project/vllm/blob/2899dca8432d40632987b0ec24253a8fe6df2710/rust/src/parser/src/unified/kimi_k3/structural_tag.rs
  - https://github.com/vllm-project/vllm/blob/2899dca8432d40632987b0ec24253a8fe6df2710/rust/src/chat/src/renderer/kimi_k3/tests.rs
  - https://github.com/vllm-project/vllm/blob/2899dca8432d40632987b0ec24253a8fe6df2710/rust/src/chat/src/renderer/kimi_k3/fixtures/tools_history_and_required_input.json
  - https://github.com/vllm-project/vllm/blob/2899dca8432d40632987b0ec24253a8fe6df2710/rust/src/server/src/routes/openai/chat_completions/convert.rs
  - https://github.com/vllm-project/vllm/blob/2899dca8432d40632987b0ec24253a8fe6df2710/rust/src/chat/src/renderer/kimi_k3/fixtures/dynamic_system_tool_declare_input.json
  - https://github.com/vllm-project/vllm/blob/2899dca8432d40632987b0ec24253a8fe6df2710/rust/src/chat/src/renderer/kimi_k3/mod.rs
state: merged
tags:
  - vllm
  - vllm-pr
  - new-architectures
title: '#50104 [Model] Add Kimi K3 support: Rust frontend [1/2]'
topic: new-architectures
url: https://github.com/vllm-project/vllm/pull/50104
created: '2026-11-11'
modified: '2026-11-11'
published: '2026-11-11'
pageLayout: default
slug: hinterland/papers/maps/prs/50104
permalink: https://aarnphm.xyz/hinterland/papers/maps/prs/50104.md
generator:
  quartz: v4.6.0
  hostedProvider: Cloudflare
  baseUrl: aarnphm.xyz
full: https://aarnphm.xyz/llms-full.txt
---
Extract the Kimi K3 Rust frontend from [#50000](https://github.com/aarnphm/aarnphm.github.io/issues/50000) into an independently reviewable stack rooted at `main`. This layer adds native XTML rendering, unified parsing, tokenizer/model selection, structured-tag support, and top-level… · full dump: [[thoughts/inference/vllm-prs/50104|#50104]]

