From 1f41de1611d9696b5c60fd509d4e651bfd03048c Mon Sep 17 00:00:00 2001
From: localai-org-maint-bot
 <306269227+localai-org-maint-bot@users.noreply.github.com>
Date: Mon, 26 Oct 2026 16:05:00 +0000
Subject: [PATCH] feat(gallery): add Nanbeige4.3 GGUF variants

Add AD-Q4_K_M and Q8_0 builds with pinned URLs and verified checksums.
Group the builds and document installation with a 32K default context.

Assisted-by: Codex:gpt-6
---
 docs/content/features/model-gallery.md | 16 +++++
 gallery/index.yaml                     | 66 +++++++++++++++++++++
 gallery/models/index.yaml              | 82 ++++++++++++++++++++++++++
 3 files changed, 164 insertions(+)

diff --git a/docs/content/features/model-gallery.md b/docs/content/features/model-gallery.md
index ad3596181e..9370d39540 100644
--- a/docs/content/features/model-gallery.md
+++ b/docs/content/features/model-gallery.md
@@ -191,6 +191,22 @@ frees. Removal then waits 30 seconds, during which you can undo it; the delete
 request is sent only when that time ends. If you leave the page during the wait,
 nothing is deleted.
 
+## Nanbeige4.3-33B
+
+[Nanbeige4.3-33B](https://huggingface.co/Nanbeige/Nanbeige4.3-33B) is a text model for reasoning, coding, and agent tasks. The gallery offers AtomicChat AD-Q4_K_M and Q8_0 GGUF builds for `llama-cpp`. Both use the embedded chat template with thinking enabled and a default context of 32,768 tokens.
+
+Install the grouped entry to let LocalAI select a build that fits the host:
+
+```bash
+local-ai models install nanbeige4.3-33b-ad-q4
+```
+
+To select Q8_0 explicitly:
+
+```bash
+local-ai models install nanbeige4.3-33b-q8
+```
+
 ## LFM2.5-24B-A2B
 
 [LiquidAI LFM2.5-24B-A2B](https://huggingface.co/LiquidAI/LFM2.5-24B-A2B)
diff --git a/gallery/index.yaml b/gallery/index.yaml
index b3d4cc2183..bec0d23182 100644
--- a/gallery/index.yaml
+++ b/gallery/index.yaml
@@ -23663,6 +23663,72 @@
       model: llama-cpp/models/Qwen_Qwen3-Next-80B-A3B-Thinking-Q4_K_M.gguf
     template:
       use_tokenizer_template: true
+- url: github:mudler/LocalAI/gallery/virtual.yaml@master
+  name: nanbeige4.3-33b-ad-q4
+  description: |
+    Nanbeige4.3-33B is a 33B dense model for reasoning, coding, and
+    agent tasks in English and Chinese. This AD-Q4_K_M GGUF build
+    uses llama.cpp with the embedded chat template and thinking enabled.
+  license: apache-2.0
+  urls:
+    - https://huggingface.co/Nanbeige/Nanbeige4.3-33B
+    - https://huggingface.co/AtomicChat/Nanbeige4.3-33B-GGUF
+  tags:
+    - llm
+    - gguf
+    - reasoning
+    - coding
+    - cpu
+    - gpu
+  files:
+    - filename: Nanbeige4.3-33B-AD-Q4_K_M.gguf
+      sha256: 409c92c4355c81928f0f2de5d49481fe7e4d0fa4d381aa28eddee0dc4b358c49
+      uri: https://huggingface.co/AtomicChat/Nanbeige4.3-33B-GGUF/resolve/b39515bab760dc8d6f7da7b4073847997971e2b7/Nanbeige4.3-33B-AD-Q4_K_M.gguf
+  overrides:
+    backend: llama-cpp
+    context_size: 32768
+    options:
+      - use_jinja:true
+    parameters:
+      model: Nanbeige4.3-33B-AD-Q4_K_M.gguf
+      temperature: 0.6
+      top_p: 0.95
+    template:
+      use_tokenizer_template: true
+  variants:
+    - model: nanbeige4.3-33b-q8
+- url: github:mudler/LocalAI/gallery/virtual.yaml@master
+  name: nanbeige4.3-33b-q8
+  description: |
+    Nanbeige4.3-33B is a 33B dense model for reasoning, coding, and
+    agent tasks in English and Chinese. This Q8_0 GGUF build
+    uses llama.cpp with the embedded chat template and thinking enabled.
+  license: apache-2.0
+  urls:
+    - https://huggingface.co/Nanbeige/Nanbeige4.3-33B
+    - https://huggingface.co/AtomicChat/Nanbeige4.3-33B-GGUF
+  tags:
+    - llm
+    - gguf
+    - reasoning
+    - coding
+    - cpu
+    - gpu
+  files:
+    - filename: Nanbeige4.3-33B-Q8_0.gguf
+      sha256: 4106cbb9ba85da731f88b7493a90343e0d8cd6c77974ed1af00a97c5262c22bc
+      uri: https://huggingface.co/AtomicChat/Nanbeige4.3-33B-GGUF/resolve/b39515bab760dc8d6f7da7b4073847997971e2b7/Nanbeige4.3-33B-Q8_0.gguf
+  overrides:
+    backend: llama-cpp
+    context_size: 32768
+    options:
+      - use_jinja:true
+    parameters:
+      model: Nanbeige4.3-33B-Q8_0.gguf
+      temperature: 0.6
+      top_p: 0.95
+    template:
+      use_tokenizer_template: true
 - url: github:mudler/LocalAI/gallery/nanbeige4.1.yaml@master
   name: nanbeige4.1-3b-q8
   description: |
diff --git a/gallery/models/index.yaml b/gallery/models/index.yaml
index 000476beb1..d3e2afd9b6 100644
--- a/gallery/models/index.yaml
+++ b/gallery/models/index.yaml
@@ -27664,6 +27664,88 @@
         model: llama-cpp/models/Qwen_Qwen3-Next-80B-A3B-Thinking-Q4_K_M.gguf
       template:
         use_tokenizer_template: true
+- kind: model
+  id: nanbeige4.3-33b-ad-q4
+  name: nanbeige4.3-33b-ad-q4
+  version: 0.0.0-gallery.b39515bab760
+  description: |
+    Nanbeige4.3-33B is a 33B dense model for reasoning, coding, and
+    agent tasks in English and Chinese. This AD-Q4_K_M GGUF build
+    uses llama.cpp with the embedded chat template and thinking enabled.
+  license: apache-2.0
+  urls:
+    - https://huggingface.co/Nanbeige/Nanbeige4.3-33B
+    - https://huggingface.co/AtomicChat/Nanbeige4.3-33B-GGUF
+  tags:
+    - llm
+    - gguf
+    - reasoning
+    - coding
+    - cpu
+    - gpu
+  channel: gallery
+  policy:
+    requireSignature: false
+  schemaVersion: 1
+  model:
+    url: github:mudler/LocalAI/gallery/virtual.yaml@master
+    files:
+      - filename: Nanbeige4.3-33B-AD-Q4_K_M.gguf
+        sha256: 409c92c4355c81928f0f2de5d49481fe7e4d0fa4d381aa28eddee0dc4b358c49
+        uri: https://huggingface.co/AtomicChat/Nanbeige4.3-33B-GGUF/resolve/b39515bab760dc8d6f7da7b4073847997971e2b7/Nanbeige4.3-33B-AD-Q4_K_M.gguf
+    overrides:
+      backend: llama-cpp
+      context_size: 32768
+      options:
+        - use_jinja:true
+      parameters:
+        model: Nanbeige4.3-33B-AD-Q4_K_M.gguf
+        temperature: 0.6
+        top_p: 0.95
+      template:
+        use_tokenizer_template: true
+    variants:
+      - model: nanbeige4.3-33b-q8
+- kind: model
+  id: nanbeige4.3-33b-q8
+  name: nanbeige4.3-33b-q8
+  version: 0.0.0-gallery.b39515bab760
+  description: |
+    Nanbeige4.3-33B is a 33B dense model for reasoning, coding, and
+    agent tasks in English and Chinese. This Q8_0 GGUF build
+    uses llama.cpp with the embedded chat template and thinking enabled.
+  license: apache-2.0
+  urls:
+    - https://huggingface.co/Nanbeige/Nanbeige4.3-33B
+    - https://huggingface.co/AtomicChat/Nanbeige4.3-33B-GGUF
+  tags:
+    - llm
+    - gguf
+    - reasoning
+    - coding
+    - cpu
+    - gpu
+  channel: gallery
+  policy:
+    requireSignature: false
+  schemaVersion: 1
+  model:
+    url: github:mudler/LocalAI/gallery/virtual.yaml@master
+    files:
+      - filename: Nanbeige4.3-33B-Q8_0.gguf
+        sha256: 4106cbb9ba85da731f88b7493a90343e0d8cd6c77974ed1af00a97c5262c22bc
+        uri: https://huggingface.co/AtomicChat/Nanbeige4.3-33B-GGUF/resolve/b39515bab760dc8d6f7da7b4073847997971e2b7/Nanbeige4.3-33B-Q8_0.gguf
+    overrides:
+      backend: llama-cpp
+      context_size: 32768
+      options:
+        - use_jinja:true
+      parameters:
+        model: Nanbeige4.3-33B-Q8_0.gguf
+        temperature: 0.6
+        top_p: 0.95
+      template:
+        use_tokenizer_template: true
 - kind: model
   id: nanbeige4.1-3b-q8
   name: nanbeige4.1-3b-q8