chore(gallery): add FrogNano Q5 and Q6 variants

Fill the quantization gap between the existing Q4 and Q8 builds. Pin both downloads to the same publisher revision and verify SHA256 values against LFS metadata and x-linked-etag headers.

Assisted-by: Codex:gpt-6

docs/content/features/model-gallery.md

@@ -23,6 +23,21 @@ GPT and text generation models might have a license which is not permissive for
 
 ![output](https://github.com/mudler/LocalAI/assets/2420543/7b16676e-d5b1-4c97-89bd-9fa5065c21ad)
 
+## FrogNano coding model
+
+`frognano-4b-2609` offers Q4_K_M, Q5_K_M, Q6_K, and Q8_0 GGUF builds
+through llama.cpp. The Q5 and Q6 variants provide intermediate sizes between
+the Q4 and Q8 builds. To select one explicitly:
+
+```bash
+local-ai models install frognano-4b-2609 --variant frognano-4b-2609-q5
+local-ai models install frognano-4b-2609 --variant frognano-4b-2609-q6
+```
+
+All builds use the embedded chat template and a 32,768-token default context.
+They are configured for text-based coding and tool use. The publisher does
+not support the checkpoint's inherited image and video components.
+
 ## How it works
 
 Open **Models** in the WebUI. It is the canonical page for a model's complete

gallery/index.yaml

@@ -69293,6 +69293,8 @@
 - name: frognano-4b-2609
   url: github:mudler/LocalAI/gallery/virtual.yaml@master
   variants:
+    - model: frognano-4b-2609-q5
+    - model: frognano-4b-2609-q6
     - model: frognano-4b-2609-q8
   urls:
     - https://huggingface.co/microsoft/FrogNano-4B-2609
@@ -69331,6 +69333,84 @@
     - filename: FrogNano-4B-2609-Q4_K_M.gguf
       uri: https://huggingface.co/bartowski/FrogNano-4B-2609-GGUF/resolve/dfbb801e04ddf045ffd245b9f1a546ed0c7d960f/FrogNano-4B-2609-Q4_K_M.gguf
       sha256: 01fba00c57e2976399ae4947f0b4fdde4623d3e410a435bf2ae39ddf164e9b0a
+- name: frognano-4b-2609-q5
+  url: github:mudler/LocalAI/gallery/virtual.yaml@master
+  urls:
+    - https://huggingface.co/microsoft/FrogNano-4B-2609
+    - https://huggingface.co/bartowski/FrogNano-4B-2609-GGUF
+  description: |
+    FrogNano is Microsoft's Qwen3.5-4B derivative for repository-level coding
+    and tool use. This Q5_K_M GGUF build uses the embedded chat template
+    with a 32K-token default context. Configured for text input; upstream
+    does not evaluate or support its inherited image and video components.
+  license: mit
+  tags:
+    - llm
+    - gguf
+    - cpu
+    - gpu
+    - coding
+    - reasoning
+  last_checked: "2026-10-07"
+  overrides:
+    backend: llama-cpp
+    context_size: 32768
+    function:
+      grammar:
+        disable: true
+    known_usecases:
+      - chat
+    options:
+      - use_jinja:true
+    parameters:
+      model: FrogNano-4B-2609-Q5_K_M.gguf
+      temperature: 0.6
+      repeat_penalty: 1.0
+    template:
+      use_tokenizer_template: true
+  files:
+    - filename: FrogNano-4B-2609-Q5_K_M.gguf
+      uri: https://huggingface.co/bartowski/FrogNano-4B-2609-GGUF/resolve/dfbb801e04ddf045ffd245b9f1a546ed0c7d960f/FrogNano-4B-2609-Q5_K_M.gguf
+      sha256: f37ff83a6229a127866c1deb414ae36f6c9c30e1b21d8e0bbe313f826cb23601
+- name: frognano-4b-2609-q6
+  url: github:mudler/LocalAI/gallery/virtual.yaml@master
+  urls:
+    - https://huggingface.co/microsoft/FrogNano-4B-2609
+    - https://huggingface.co/bartowski/FrogNano-4B-2609-GGUF
+  description: |
+    FrogNano is Microsoft's Qwen3.5-4B derivative for repository-level coding
+    and tool use. This Q6_K GGUF build uses the embedded chat template
+    with a 32K-token default context. Configured for text input; upstream
+    does not evaluate or support its inherited image and video components.
+  license: mit
+  tags:
+    - llm
+    - gguf
+    - cpu
+    - gpu
+    - coding
+    - reasoning
+  last_checked: "2026-10-07"
+  overrides:
+    backend: llama-cpp
+    context_size: 32768
+    function:
+      grammar:
+        disable: true
+    known_usecases:
+      - chat
+    options:
+      - use_jinja:true
+    parameters:
+      model: FrogNano-4B-2609-Q6_K.gguf
+      temperature: 0.6
+      repeat_penalty: 1.0
+    template:
+      use_tokenizer_template: true
+  files:
+    - filename: FrogNano-4B-2609-Q6_K.gguf
+      uri: https://huggingface.co/bartowski/FrogNano-4B-2609-GGUF/resolve/dfbb801e04ddf045ffd245b9f1a546ed0c7d960f/FrogNano-4B-2609-Q6_K.gguf
+      sha256: 855278f785dc23fc4c150bf74e20ca6ed53036c06fb04abfd0ddc60d02116b6a
 - name: frognano-4b-2609-q8
   url: github:mudler/LocalAI/gallery/virtual.yaml@master
   urls: