chore(gallery): add AstaBrief-8B GGUF variants

Offer Q4_K_M and Q8_0 builds for cited research reports from supplied
literature excerpts. Use the embedded Qwen3 template and document the
upstream prompt. Pin both downloads and cross-check their SHA256 values
against Hugging Face LFS metadata and download headers.

Assisted-by: Codex:gpt-6

docs/content/features/model-gallery.md

@@ -153,6 +153,28 @@ MTP speculative decoding is not enabled by these entries.
 The weights use [PolyForm Small Business 1.0.0 with a personal-use grant](https://huggingface.co/bottlecapai/ThinkingCap-Qwen3.8-27B-GGUF/blob/main/LICENSE).
 Review that license for permitted use.
 
+## AstaBrief-8B
+
+[AstaBrief-8B](https://huggingface.co/allenai/AstaBrief_8B) is Ai2's Qwen3-8B
+fine-tune for writing cited reports from scientific literature excerpts.
+Install the llama.cpp GGUF entry with:
+
+```bash
+local-ai models install astabrief-8b-q4
+```
+
+The entry offers Q4_K_M and Q8_0 builds. To select a specific build, use
+`--variant astabrief-8b-q4` or `--variant astabrief-8b-q8` during installation.
+Both use the embedded chat template, a 32,768-token context, temperature `0.7`,
+and top-p `0.95`.
+
+Use the upstream [research prompt](https://huggingface.co/datasets/allenai/AstaBrief_prompts/blob/main/sft_prompt.txt).
+Replace `[QUERY]` with your question and `[SECTION_REFERENCES]` with the cited
+excerpts, then send the completed prompt as a user message to
+`/v1/chat/completions`. The model writes from the material you supply; this
+gallery entry does not search for or retrieve papers. The upstream example
+uses `max_tokens: 4096` for the report.
+
 ## Hemmingway-1
 
 Install `hemmingway-1` for English text generation with llama.cpp. The gallery groups its Q4_K_M and Q8_0 builds as variants.

gallery/index.yaml

@@ -27226,6 +27226,92 @@
     - filename: Zero-Qwen3-8B-OpenZero-Q5_K_M.gguf
       sha256: 390464f750b5cb53da298848adc05839c1fd40404a74cd5f800cad9612d17d59
       uri: huggingface://shafire/Zero-Qwen3-8B-OpenZero-GGUF/Zero-Qwen3-8B-OpenZero-Q5_K_M.gguf
+- name: astabrief-8b-q4
+  url: github:mudler/LocalAI/gallery/virtual.yaml@master
+  urls:
+    - https://huggingface.co/allenai/AstaBrief_8B
+    - https://huggingface.co/mradermacher/AstaBrief_8B-GGUF
+  description: |
+    AstaBrief-8B is Ai2's Qwen3-8B fine-tune for writing cited research reports
+    from a question and supplied scientific literature excerpts. This Q4_K_M
+    GGUF build uses llama.cpp and the embedded chat template. Provide the
+    upstream research prompt and references in the user message; the model
+    does not retrieve papers itself.
+  license: apache-2.0
+  tags:
+    - llm
+    - gguf
+    - cpu
+    - gpu
+    - qwen3
+    - research
+    - en
+  variants:
+    - model: astabrief-8b-q8
+  overrides:
+    backend: llama-cpp
+    known_usecases:
+      - chat
+    context_size: 32768
+    mmap: true
+    f16: true
+    options:
+      - use_jinja:true
+    template:
+      use_tokenizer_template: true
+    parameters:
+      model: llama-cpp/models/astabrief-8b/AstaBrief_8B.Q4_K_M.gguf
+      temperature: 0.7
+      top_p: 0.95
+      top_k: 0
+    stopwords:
+      - <|im_end|>
+  files:
+    - filename: llama-cpp/models/astabrief-8b/AstaBrief_8B.Q4_K_M.gguf
+      uri: https://huggingface.co/mradermacher/AstaBrief_8B-GGUF/resolve/a40b9eb85c6e349898737726b4b9ee69688b63a3/AstaBrief_8B.Q4_K_M.gguf
+      sha256: 1df1265b9715fc5118134010abcd3d297fd4be08a9b2294c7a467c79f161ccce
+- name: astabrief-8b-q8
+  url: github:mudler/LocalAI/gallery/virtual.yaml@master
+  urls:
+    - https://huggingface.co/allenai/AstaBrief_8B
+    - https://huggingface.co/mradermacher/AstaBrief_8B-GGUF
+  description: |
+    AstaBrief-8B is Ai2's Qwen3-8B fine-tune for writing cited research reports
+    from a question and supplied scientific literature excerpts. This Q8_0
+    GGUF build uses llama.cpp and the embedded chat template. Provide the
+    upstream research prompt and references in the user message; the model
+    does not retrieve papers itself.
+  license: apache-2.0
+  tags:
+    - llm
+    - gguf
+    - cpu
+    - gpu
+    - qwen3
+    - research
+    - en
+  overrides:
+    backend: llama-cpp
+    known_usecases:
+      - chat
+    context_size: 32768
+    mmap: true
+    f16: true
+    options:
+      - use_jinja:true
+    template:
+      use_tokenizer_template: true
+    parameters:
+      model: llama-cpp/models/astabrief-8b/AstaBrief_8B.Q8_0.gguf
+      temperature: 0.7
+      top_p: 0.95
+      top_k: 0
+    stopwords:
+      - <|im_end|>
+  files:
+    - filename: llama-cpp/models/astabrief-8b/AstaBrief_8B.Q8_0.gguf
+      uri: https://huggingface.co/mradermacher/AstaBrief_8B-GGUF/resolve/a40b9eb85c6e349898737726b4b9ee69688b63a3/AstaBrief_8B.Q8_0.gguf
+      sha256: c199cbac2cccdbe16e6ca3aed60c9868216ed1ed90a157f2451de37bba074d2a
 - name: qwen3-8b
   url: github:mudler/LocalAI/gallery/qwen3.yaml@master
   urls: