feat(gallery): add Genesis Hermes Final variants

Add APEX, Compact, Q8_K_P, and MTP builds with the F16 projector. Pin downloads to one HF revision and verify SHA256 values against LFS metadata and download headers. Document installation and context. Assisted-by: Codex:gpt-6
2026-09-25 07:34:58 -04:00 · 2026-09-14 00:08:16 +00:00
1 parent 3dfb34376a
commit ba2a61f205
@@ -821,3 +821,27 @@ is still running is reported as queued until the installer picks it up:
 A job ID is queryable from the moment `/models/apply` returns it, so a `404`/`500`
 from this endpoint means the ID is genuinely unknown rather than merely waiting
 its turn.
+
+### Genesis Hermes Final
+
+Install the Qwen3.6-35B-A3B Genesis Hermes Final model for text chat,
+function calling, and image input:
+
+```bash
+local-ai models install qwen3.6-35b-a3b-genesis-hermes-final
+```
+
+The gallery offers APEX, APEX Compact, and Q8_K_P builds for llama.cpp.
+APEX and APEX Compact also have variants with multi-token prediction (MTP)
+enabled. Each build includes the F16 vision projector and uses the embedded
+Jinja chat template. LocalAI selects a variant according to available memory
+and serving features. To select the plain APEX build explicitly:
+
+```bash
+local-ai models install --variant qwen3.6-35b-a3b-genesis-hermes-final qwen3.6-35b-a3b-genesis-hermes-final
+```
+
+These entries set a 131,072-token context, following the
+[model card's guidance](https://huggingface.co/LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-Final-GGUF)
+for thinking mode. This context requires additional memory beyond the weights.
+The uncensored model uses the Apache-2.0 license.
@@ -7327,6 +7327,244 @@
    - filename: llama-cpp/mmproj/Hermes3.6-35B-A3B-Uncensored-Genesis-V7/mmproj-Hermes3.6-35B-A3B-Uncensored-Genesis-F16.gguf
      uri: huggingface://LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-V7-GGUF/mmproj-Hermes3.6-35B-A3B-Uncensored-Genesis-F16.gguf
      sha256: f7197461d8581cd9be42384d4afe3f851d0844f6d72846d5c97e3dcc931b3cbc
+- &qwen3-6-35b-a3b-genesis-hermes-final
+  name: "qwen3.6-35b-a3b-genesis-hermes-final"
+  url: "github:mudler/LocalAI/gallery/virtual.yaml@master"
+  variants:
+    - model: qwen3.6-35b-a3b-genesis-hermes-final-apex-compact
+    - model: qwen3.6-35b-a3b-genesis-hermes-final-mtp-apex
+    - model: qwen3.6-35b-a3b-genesis-hermes-final-mtp-apex-compact
+    - model: qwen3.6-35b-a3b-genesis-hermes-final-q8-k-p
+  urls:
+    - https://huggingface.co/HauhauCS/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive
+    - https://huggingface.co/LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-Final-GGUF
+  description: |
+    Qwen3.6-35B-A3B Genesis Hermes Final is a multimodal mixture-of-experts
+    model with 35B total parameters and about 3B active per token. This
+    uncensored derivative combines the HauhauCS base with Hermes
+    function-calling data and the author's Genesis weight processing.
+    This build uses APEX and includes the F16 vision projector.
+  license: "apache-2.0"
+  tags:
+    - llm
+    - gguf
+    - vision
+    - multimodal
+    - reasoning
+    - coding
+    - uncensored
+    - moe
+  last_checked: "2026-09-14"
+  overrides:
+    backend: llama-cpp
+    context_size: 131072
+    function:
+      automatic_tool_parsing_fallback: true
+      grammar:
+        disable: true
+    known_usecases:
+      - chat
+      - vision
+    mmproj: llama-cpp/mmproj/Hermes3.6-35B-A3B-Uncensored-Genesis-Final/mmproj-Hermes3.6-35B-A3B-Uncensored-Genesis-Final-F16.gguf
+    options:
+      - use_jinja:true
+    parameters:
+      min_p: 0
+      model: llama-cpp/models/Hermes3.6-35B-A3B-Uncensored-Genesis-Final-APEX/Hermes3.6-35B-A3B-Uncensored-Genesis-Final-APEX.gguf
+      repeat_penalty: 1
+      temperature: 0.6
+      top_k: 20
+      top_p: 1
+    template:
+      use_tokenizer_template: true
+  files:
+    - filename: llama-cpp/models/Hermes3.6-35B-A3B-Uncensored-Genesis-Final-APEX/Hermes3.6-35B-A3B-Uncensored-Genesis-Final-APEX.gguf
+      sha256: caefdcec9364e903fc7edac90c39df257f1d6671b5fc389a3cb7dc8ce8204a3b
+      uri: https://huggingface.co/LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-Final-GGUF/resolve/d0cf3294f07f2c422f0cf793a75fa48f61d48931/Hermes3.6-35B-A3B-Uncensored-Genesis-Final-APEX.gguf
+    - filename: llama-cpp/mmproj/Hermes3.6-35B-A3B-Uncensored-Genesis-Final/mmproj-Hermes3.6-35B-A3B-Uncensored-Genesis-Final-F16.gguf
+      sha256: 5129bb5eb19e4346c0f2071f1ce8e1b0a076e0ab08d57a77d6033ce01235252c
+      uri: https://huggingface.co/LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-Final-GGUF/resolve/d0cf3294f07f2c422f0cf793a75fa48f61d48931/mmproj-Hermes3.6-35B-A3B-Uncensored-Genesis-Final-F16.gguf
+- !!merge <<: *qwen3-6-35b-a3b-genesis-hermes-final
+  name: "qwen3.6-35b-a3b-genesis-hermes-final-apex-compact"
+  variants: []
+  description: |
+    Qwen3.6-35B-A3B Genesis Hermes Final is a multimodal mixture-of-experts
+    model with 35B total parameters and about 3B active per token. This
+    uncensored derivative combines the HauhauCS base with Hermes
+    function-calling data and the author's Genesis weight processing.
+    This build uses APEX Compact and includes the F16 vision projector.
+  overrides:
+    backend: llama-cpp
+    context_size: 131072
+    function:
+      automatic_tool_parsing_fallback: true
+      grammar:
+        disable: true
+    known_usecases:
+      - chat
+      - vision
+    mmproj: llama-cpp/mmproj/Hermes3.6-35B-A3B-Uncensored-Genesis-Final/mmproj-Hermes3.6-35B-A3B-Uncensored-Genesis-Final-F16.gguf
+    options:
+      - use_jinja:true
+    parameters:
+      min_p: 0
+      model: llama-cpp/models/Hermes3.6-35B-A3B-Uncensored-Genesis-Final-APEX-Compact/Hermes3.6-35B-A3B-Uncensored-Genesis-Final-APEX-Compact.gguf
+      repeat_penalty: 1
+      temperature: 0.6
+      top_k: 20
+      top_p: 1
+    template:
+      use_tokenizer_template: true
+  files:
+    - filename: llama-cpp/models/Hermes3.6-35B-A3B-Uncensored-Genesis-Final-APEX-Compact/Hermes3.6-35B-A3B-Uncensored-Genesis-Final-APEX-Compact.gguf
+      sha256: c55f3a0acc94d28a192cf1262a9cc5dbdbe15ce7bad75894ce2b0d7e8b66ddb8
+      uri: https://huggingface.co/LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-Final-GGUF/resolve/d0cf3294f07f2c422f0cf793a75fa48f61d48931/Hermes3.6-35B-A3B-Uncensored-Genesis-Final-APEX-Compact.gguf
+    - filename: llama-cpp/mmproj/Hermes3.6-35B-A3B-Uncensored-Genesis-Final/mmproj-Hermes3.6-35B-A3B-Uncensored-Genesis-Final-F16.gguf
+      sha256: 5129bb5eb19e4346c0f2071f1ce8e1b0a076e0ab08d57a77d6033ce01235252c
+      uri: https://huggingface.co/LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-Final-GGUF/resolve/d0cf3294f07f2c422f0cf793a75fa48f61d48931/mmproj-Hermes3.6-35B-A3B-Uncensored-Genesis-Final-F16.gguf
+- !!merge <<: *qwen3-6-35b-a3b-genesis-hermes-final
+  name: "qwen3.6-35b-a3b-genesis-hermes-final-mtp-apex"
+  variants: []
+  description: |
+    Qwen3.6-35B-A3B Genesis Hermes Final is a multimodal mixture-of-experts
+    model with 35B total parameters and about 3B active per token. This
+    uncensored derivative combines the HauhauCS base with Hermes
+    function-calling data and the author's Genesis weight processing.
+    This build uses APEX and includes the F16 vision projector.
+    Native multi-token prediction is enabled for speculative decoding.
+  tags:
+    - llm
+    - gguf
+    - vision
+    - multimodal
+    - reasoning
+    - coding
+    - uncensored
+    - moe
+    - mtp
+  overrides:
+    backend: llama-cpp
+    context_size: 131072
+    function:
+      automatic_tool_parsing_fallback: true
+      grammar:
+        disable: true
+    known_usecases:
+      - chat
+      - vision
+    mmproj: llama-cpp/mmproj/Hermes3.6-35B-A3B-Uncensored-Genesis-Final/mmproj-Hermes3.6-35B-A3B-Uncensored-Genesis-Final-F16.gguf
+    options:
+      - use_jinja:true
+      - spec_type:draft-mtp
+      - spec_n_max:6
+      - spec_p_min:0.75
+    parameters:
+      min_p: 0
+      model: llama-cpp/models/Hermes3.6-35B-A3B-Uncensored-Genesis-Final-MTP-APEX/Hermes3.6-35B-A3B-Uncensored-Genesis-Final-MTP-APEX.gguf
+      repeat_penalty: 1
+      temperature: 0.6
+      top_k: 20
+      top_p: 1
+    template:
+      use_tokenizer_template: true
+  files:
+    - filename: llama-cpp/models/Hermes3.6-35B-A3B-Uncensored-Genesis-Final-MTP-APEX/Hermes3.6-35B-A3B-Uncensored-Genesis-Final-MTP-APEX.gguf
+      sha256: 01e16a28774e4f075a2bfd92ff889cb7b004e37511e68399caf39a3f4b195722
+      uri: https://huggingface.co/LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-Final-GGUF/resolve/d0cf3294f07f2c422f0cf793a75fa48f61d48931/Hermes3.6-35B-A3B-Uncensored-Genesis-Final-MTP-APEX.gguf
+    - filename: llama-cpp/mmproj/Hermes3.6-35B-A3B-Uncensored-Genesis-Final/mmproj-Hermes3.6-35B-A3B-Uncensored-Genesis-Final-F16.gguf
+      sha256: 5129bb5eb19e4346c0f2071f1ce8e1b0a076e0ab08d57a77d6033ce01235252c
+      uri: https://huggingface.co/LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-Final-GGUF/resolve/d0cf3294f07f2c422f0cf793a75fa48f61d48931/mmproj-Hermes3.6-35B-A3B-Uncensored-Genesis-Final-F16.gguf
+- !!merge <<: *qwen3-6-35b-a3b-genesis-hermes-final
+  name: "qwen3.6-35b-a3b-genesis-hermes-final-mtp-apex-compact"
+  variants: []
+  description: |
+    Qwen3.6-35B-A3B Genesis Hermes Final is a multimodal mixture-of-experts
+    model with 35B total parameters and about 3B active per token. This
+    uncensored derivative combines the HauhauCS base with Hermes
+    function-calling data and the author's Genesis weight processing.
+    This build uses APEX Compact and includes the F16 vision projector.
+    Native multi-token prediction is enabled for speculative decoding.
+  tags:
+    - llm
+    - gguf
+    - vision
+    - multimodal
+    - reasoning
+    - coding
+    - uncensored
+    - moe
+    - mtp
+  overrides:
+    backend: llama-cpp
+    context_size: 131072
+    function:
+      automatic_tool_parsing_fallback: true
+      grammar:
+        disable: true
+    known_usecases:
+      - chat
+      - vision
+    mmproj: llama-cpp/mmproj/Hermes3.6-35B-A3B-Uncensored-Genesis-Final/mmproj-Hermes3.6-35B-A3B-Uncensored-Genesis-Final-F16.gguf
+    options:
+      - use_jinja:true
+      - spec_type:draft-mtp
+      - spec_n_max:6
+      - spec_p_min:0.75
+    parameters:
+      min_p: 0
+      model: llama-cpp/models/Hermes3.6-35B-A3B-Uncensored-Genesis-Final-MTP-APEX-Compact/Hermes3.6-35B-A3B-Uncensored-Genesis-Final-MTP-APEX-Compact.gguf
+      repeat_penalty: 1
+      temperature: 0.6
+      top_k: 20
+      top_p: 1
+    template:
+      use_tokenizer_template: true
+  files:
+    - filename: llama-cpp/models/Hermes3.6-35B-A3B-Uncensored-Genesis-Final-MTP-APEX-Compact/Hermes3.6-35B-A3B-Uncensored-Genesis-Final-MTP-APEX-Compact.gguf
+      sha256: fe921f9f7d3f48f9fbda1f28bb3e96824db051958cbc6e44dd0e5bebc690829a
+      uri: https://huggingface.co/LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-Final-GGUF/resolve/d0cf3294f07f2c422f0cf793a75fa48f61d48931/Hermes3.6-35B-A3B-Uncensored-Genesis-Final-MTP-APEX-Compact.gguf
+    - filename: llama-cpp/mmproj/Hermes3.6-35B-A3B-Uncensored-Genesis-Final/mmproj-Hermes3.6-35B-A3B-Uncensored-Genesis-Final-F16.gguf
+      sha256: 5129bb5eb19e4346c0f2071f1ce8e1b0a076e0ab08d57a77d6033ce01235252c
+      uri: https://huggingface.co/LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-Final-GGUF/resolve/d0cf3294f07f2c422f0cf793a75fa48f61d48931/mmproj-Hermes3.6-35B-A3B-Uncensored-Genesis-Final-F16.gguf
+- !!merge <<: *qwen3-6-35b-a3b-genesis-hermes-final
+  name: "qwen3.6-35b-a3b-genesis-hermes-final-q8-k-p"
+  variants: []
+  description: |
+    Qwen3.6-35B-A3B Genesis Hermes Final is a multimodal mixture-of-experts
+    model with 35B total parameters and about 3B active per token. This
+    uncensored derivative combines the HauhauCS base with Hermes
+    function-calling data and the author's Genesis weight processing.
+    This build uses Q8_K_P and includes the F16 vision projector.
+  overrides:
+    backend: llama-cpp
+    context_size: 131072
+    function:
+      automatic_tool_parsing_fallback: true
+      grammar:
+        disable: true
+    known_usecases:
+      - chat
+      - vision
+    mmproj: llama-cpp/mmproj/Hermes3.6-35B-A3B-Uncensored-Genesis-Final/mmproj-Hermes3.6-35B-A3B-Uncensored-Genesis-Final-F16.gguf
+    options:
+      - use_jinja:true
+    parameters:
+      min_p: 0
+      model: llama-cpp/models/Hermes3.6-35B-A3B-Uncensored-Genesis-Final-Q8_K_P/Hermes3.6-35B-A3B-Uncensored-Genesis-Final-Q8_K_P.gguf
+      repeat_penalty: 1
+      temperature: 0.6
+      top_k: 20
+      top_p: 1
+    template:
+      use_tokenizer_template: true
+  files:
+    - filename: llama-cpp/models/Hermes3.6-35B-A3B-Uncensored-Genesis-Final-Q8_K_P/Hermes3.6-35B-A3B-Uncensored-Genesis-Final-Q8_K_P.gguf
+      sha256: e1ac7246dd3ebfb036732874dc57acbacc8b0d7d02df4417e2ac0d1f376e8d1d
+      uri: https://huggingface.co/LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-Final-GGUF/resolve/d0cf3294f07f2c422f0cf793a75fa48f61d48931/Hermes3.6-35B-A3B-Uncensored-Genesis-Final-Q8_K_P.gguf
+    - filename: llama-cpp/mmproj/Hermes3.6-35B-A3B-Uncensored-Genesis-Final/mmproj-Hermes3.6-35B-A3B-Uncensored-Genesis-Final-F16.gguf
+      sha256: 5129bb5eb19e4346c0f2071f1ce8e1b0a076e0ab08d57a77d6033ce01235252c
+      uri: https://huggingface.co/LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-Final-GGUF/resolve/d0cf3294f07f2c422f0cf793a75fa48f61d48931/mmproj-Hermes3.6-35B-A3B-Uncensored-Genesis-Final-F16.gguf
+
 - &kat-coder-v2-5-dev
  name: "kat-coder-v2.5-dev"
  variants: