@bendyline/gilde 0.1.29 → 0.1.31

package/data/chat-models/ll/llama3.2-3b-q4/versions/1.1.1/manifest.json

@@ -0,0 +1,56 @@
+{
+  "schemaVersion": 1,
+  "version": "1.1.1",
+  "releasedAt": "2026-08-15T00:00:00Z",
+  "approxSizeBytes": 2019377696,
+  "ollama": {
+    "tag": "llama3.2"
+  },
+  "llamaCpp": {
+    "huggingfaceRepo": "bartowski/Llama-3.2-3B-Instruct-GGUF",
+    "revision": "5ab33fa94d1d04e903623ae72c95d1696f09f9e8",
+    "filename": "Llama-3.2-3B-Instruct-Q4_K_M.gguf",
+    "sha256": "6c1a2b41161032677be168d354123594c0e6e67d2b9227c84f296ad037c728ff",
+    "approxSizeBytes": 2019377696,
+    "quantization": "Q4_K_M"
+  },
+  "mlx": {
+    "huggingfaceRepo": "mlx-community/Llama-3.2-3B-Instruct-4bit",
+    "revision": "7f0dc925e0d0afb0322d96f9255cfddf2ba5636e",
+    "quantization": "4bit",
+    "approxSizeBytes": 1824807894,
+    "files": [
+      {
+        "name": "config.json",
+        "sha256": "c546925585e48f43890d9dc5150df4fec73dd3780d92961c5ace451934cc4cd6",
+        "sizeBytes": 1122
+      },
+      {
+        "name": "model.safetensors",
+        "sha256": "d75e1ee0ea653cc5b76191ec934c7c0d568e94d4e47846619f1f4bc715b7b265",
+        "sizeBytes": 1807496278
+      },
+      {
+        "name": "model.safetensors.index.json",
+        "sha256": "2ef31fa0b9dcda01f87835851d5e1d5a39ab6258ae618af6ea864c3747e556e4",
+        "sizeBytes": 45720
+      },
+      {
+        "name": "special_tokens_map.json",
+        "sha256": "6f38c73729248f6c127296386e3cdde96e254636cc58b4169d3fd32328d9a8ec",
+        "sizeBytes": 296
+      },
+      {
+        "name": "tokenizer.json",
+        "sha256": "6b9e4e7fb171f92fd137b777cc2714bf87d11576700a1dcd7a399e7bbe39537b",
+        "sizeBytes": 17209920
+      },
+      {
+        "name": "tokenizer_config.json",
+        "sha256": "022d5ae3df4737998ab97d8f31ac2bcb4c06dd8ebe5a8aba2b4aceef1e5ea7d3",
+        "sizeBytes": 54558
+      }
+    ]
+  },
+  "minGezelVersion": "1.26227"
+}

package/data/chat-models/mi/mistral-7b-q4/manifest.json

@@ -13,8 +13,8 @@
     "name": "Mistral AI",
     "url": "https://ollama.com/library/mistral"
   },
-  "version": "1.1.0",
-  "updatedAt": "2026-04-26T00:00:00Z",
+  "version": "1.1.1",
+  "updatedAt": "2026-08-15T00:00:00Z",
   "license": "Apache-2.0",
   "licenseClass": "open",
   "licenseShortName": "Apache 2.0",
@@ -46,8 +46,7 @@
     "filename": "Mistral-7B-Instruct-v0.3-Q4_K_M.gguf",
     "sha256": "1270d22c0fbb3d092fb725d4d96c457b7b687a5f5a715abe1e818da303e562b6",
     "approxSizeBytes": 4372812000,
-    "quantization": "Q4_K_M",
-    "residentBytes": 5247374400
+    "quantization": "Q4_K_M"
   },
   "mlx": {
     "huggingfaceRepo": "mlx-community/Mistral-7B-Instruct-v0.3-4bit",
@@ -90,8 +89,7 @@
         "sha256": "e8fbcfef34d855dba2c7f1d91a2c5470cf493d00ecfd06488ce21e45325bc8ae",
         "sizeBytes": 138039
       }
-    ],
-    "residentBytes": 5304286939
+    ]
   },
   "tuning": {
     "engine": {

package/data/chat-models/mi/mistral-7b-q4/versions/1.1.1/manifest.json

@@ -0,0 +1,61 @@
+{
+  "schemaVersion": 1,
+  "version": "1.1.1",
+  "releasedAt": "2026-08-15T00:00:00Z",
+  "approxSizeBytes": 4372812000,
+  "ollama": {
+    "tag": "mistral:7b"
+  },
+  "llamaCpp": {
+    "huggingfaceRepo": "bartowski/Mistral-7B-Instruct-v0.3-GGUF",
+    "revision": "61fd4167fff3ab01ee1cfe0da183fa27a944db48",
+    "filename": "Mistral-7B-Instruct-v0.3-Q4_K_M.gguf",
+    "sha256": "1270d22c0fbb3d092fb725d4d96c457b7b687a5f5a715abe1e818da303e562b6",
+    "approxSizeBytes": 4372812000,
+    "quantization": "Q4_K_M"
+  },
+  "mlx": {
+    "huggingfaceRepo": "mlx-community/Mistral-7B-Instruct-v0.3-4bit",
+    "revision": "a4b8f870474b0eb527f466a03fbc187830d271f5",
+    "quantization": "4bit",
+    "approxSizeBytes": 4080220722,
+    "files": [
+      {
+        "name": "config.json",
+        "sha256": "efc145d2f53d6d6ef2639be0556b4ecf6faac6cf8a7199ad2b7bb5ebcf443bdd",
+        "sizeBytes": 721
+      },
+      {
+        "name": "model.safetensors",
+        "sha256": "d7182d886f93a9efa73d6d556ee33ffe839a4e09e15965eeef46f295ab5b503d",
+        "sizeBytes": 4077480215
+      },
+      {
+        "name": "model.safetensors.index.json",
+        "sha256": "525940738b16d210fdfd692a24752d050c15a50cbe208dd82b06204d640d687b",
+        "sizeBytes": 52381
+      },
+      {
+        "name": "special_tokens_map.json",
+        "sha256": "6fa06efa2785e450051989a6f8fb4416b10149ded485ddd3f127a40734f5cfd0",
+        "sizeBytes": 414
+      },
+      {
+        "name": "tokenizer.json",
+        "sha256": "e553af6fff7d7ad76e830608b218c5c0b0822998d5a1a96099a74cd3c1cb1a49",
+        "sizeBytes": 1961548
+      },
+      {
+        "name": "tokenizer.model",
+        "sha256": "37f00374dea48658ee8f5d0f21895b9bc55cb0103939607c8185bfd1c6ca1f89",
+        "sizeBytes": 587404
+      },
+      {
+        "name": "tokenizer_config.json",
+        "sha256": "e8fbcfef34d855dba2c7f1d91a2c5470cf493d00ecfd06488ce21e45325bc8ae",
+        "sizeBytes": 138039
+      }
+    ]
+  },
+  "minGezelVersion": "1.26227"
+}

package/data/chat-models/mi/mistral-medium-3.5-128b-q4/manifest.json

@@ -16,8 +16,8 @@
     "name": "Mistral AI",
     "url": "https://huggingface.co/mistralai/Mistral-Medium-3.5-128B"
   },
-  "version": "1.0.0",
-  "updatedAt": "2026-05-22T00:00:00Z",
+  "version": "1.0.1",
+  "updatedAt": "2026-08-15T00:00:00Z",
   "license": "MIT-Modified",
   "licenseClass": "custom-restricted",
   "licenseShortName": "MIT-Modified",

package/data/chat-models/mi/mistral-medium-3.5-128b-q4/versions/1.0.1/manifest.json

@@ -0,0 +1,153 @@
+{
+  "schemaVersion": 1,
+  "version": "1.0.1",
+  "releasedAt": "2026-08-15T00:00:00Z",
+  "approxSizeBytes": 74897139136,
+  "llamaCpp": {
+    "huggingfaceRepo": "unsloth/Mistral-Medium-3.5-128B-GGUF",
+    "revision": "c8f5b1477e1b22cd2d819157d450f001f7047298",
+    "shards": [
+      {
+        "name": "Q4_K_M/Mistral-Medium-3.5-128B-Q4_K_M-00001-of-00003.gguf",
+        "sha256": "94bf2dd90d80f41b1dfb1f94b569d11fd78fd302a79ae9fdf828f8082d1dfeaa",
+        "sizeBytes": 7875456
+      },
+      {
+        "name": "Q4_K_M/Mistral-Medium-3.5-128B-Q4_K_M-00002-of-00003.gguf",
+        "sha256": "8396392f4b9eafac69e59355b569b6603f42d103505988e26939db7f5bfecb95",
+        "sizeBytes": 49927945056
+      },
+      {
+        "name": "Q4_K_M/Mistral-Medium-3.5-128B-Q4_K_M-00003-of-00003.gguf",
+        "sha256": "3f4ddc0284d522e6f1a020eaa731eaf851d6708a26431773de7bba504227c2e5",
+        "sizeBytes": 24961318624
+      }
+    ],
+    "approxSizeBytes": 74897139136,
+    "quantization": "Q4_K_M"
+  },
+  "mlx": {
+    "huggingfaceRepo": "mlx-community/Mistral-Medium-3.5-128B-4bit",
+    "revision": "5a7b6b72ca9eb2324a90bdb5e6079789dae68348",
+    "quantization": "4bit",
+    "approxSizeBytes": 78019864102,
+    "files": [
+      {
+        "name": "chat_template.jinja",
+        "sha256": "07ef40663bd181457c9c2a823296c83b33dc4f09e5d5f11478c219f7fb4f5633",
+        "sizeBytes": 13479
+      },
+      {
+        "name": "config.json",
+        "sha256": "915a6627d7cef04adf8efbb7c8befa1f09833d038eeddd80958aa88f5ae5dabb",
+        "sizeBytes": 2127
+      },
+      {
+        "name": "consolidated.safetensors.index.json",
+        "sha256": "abf847cd7a832d633fc800945dabf54fad3fca04a853c7d53eefa1e9b90fd8d0",
+        "sizeBytes": 217908
+      },
+      {
+        "name": "generation_config.json",
+        "sha256": "a107bf29bb85bd142694b293f41f5984176c598c8c3a3bc47f1f5ac8008bf89e",
+        "sizeBytes": 131
+      },
+      {
+        "name": "model-00001-of-00015.safetensors",
+        "sha256": "595e23a7aaf89c392165e653cb8d4cf862aa596f6cbfb1d03ed671330f72d0de",
+        "sizeBytes": 5356508226
+      },
+      {
+        "name": "model-00002-of-00015.safetensors",
+        "sha256": "59db0909274c1464a058fb6d98ba5ed39280bf96ed34f81867b27351e1c9154d",
+        "sizeBytes": 5368448119
+      },
+      {
+        "name": "model-00003-of-00015.safetensors",
+        "sha256": "91e857ebe77e1020f7e90742ee643290f71a67cccd56ba81250daeb2e2876e64",
+        "sizeBytes": 5263167745
+      },
+      {
+        "name": "model-00004-of-00015.safetensors",
+        "sha256": "dfa77b77d40ff44aa27a8be3d1a611e177a9e29b7fcf47d99959643b115bb435",
+        "sizeBytes": 5252157665
+      },
+      {
+        "name": "model-00005-of-00015.safetensors",
+        "sha256": "0b62abf4d755c9a882e513f74cfebcffb7cce94fe6746aff4912eb411174a962",
+        "sizeBytes": 5365403842
+      },
+      {
+        "name": "model-00006-of-00015.safetensors",
+        "sha256": "fc241be9252defbe6bd44c1ec3c2c1cc7e3cbfef3c710025944aa45c91e63477",
+        "sizeBytes": 5351247306
+      },
+      {
+        "name": "model-00007-of-00015.safetensors",
+        "sha256": "95ba57e62c4979b6dcc23dc4d4d199c7ac6c71b5f9a96d20ebf99e231202fe36",
+        "sizeBytes": 5252108242
+      },
+      {
+        "name": "model-00008-of-00015.safetensors",
+        "sha256": "574edc0a63b23874258610a132a69c648520dbd8db2a6a68929ead73872c8d67",
+        "sizeBytes": 5252157663
+      },
+      {
+        "name": "model-00009-of-00015.safetensors",
+        "sha256": "dab1076222d40b7777d8fe09285cb804188ad8a10af1d5232ba19b842d34cec3",
+        "sizeBytes": 5252157629
+      },
+      {
+        "name": "model-00010-of-00015.safetensors",
+        "sha256": "1c7836f1b077dbffa6a54a3458dd66879f343f4f8e648d74e7f21e2642ecf48c",
+        "sizeBytes": 5365403830
+      },
+      {
+        "name": "model-00011-of-00015.safetensors",
+        "sha256": "13402c3ceccea8af794bec72d841885c7f63a7fba534a1b9c6d7f722d028c7d5",
+        "sizeBytes": 5351247296
+      },
+      {
+        "name": "model-00012-of-00015.safetensors",
+        "sha256": "b2f988520b06cc34a9f9b0c4f71b1cc0fbd465b7c45a20ead3911fe1ca434d5a",
+        "sizeBytes": 5252108278
+      },
+      {
+        "name": "model-00013-of-00015.safetensors",
+        "sha256": "97f3bb3f7b2486db4d4912d5a3620b5ae8d34b8f55ce23c17b7a553c19dc6a1f",
+        "sizeBytes": 5252157667
+      },
+      {
+        "name": "model-00014-of-00015.safetensors",
+        "sha256": "7a1f12144046d582ff6997eeefd29f025060586f5cf23880be3416bb630adce5",
+        "sizeBytes": 5252157631
+      },
+      {
+        "name": "model-00015-of-00015.safetensors",
+        "sha256": "e03cd427f534273d586bad3d9f7ff24d3eb838572dc5be7bc4bc98cdeba86d92",
+        "sizeBytes": 3815843471
+      },
+      {
+        "name": "model.safetensors.index.json",
+        "sha256": "7dd3ad9ef87bae38057d528bd92342bd6dadae4112113c2b55c8ee6ef67a85b4",
+        "sizeBytes": 256576
+      },
+      {
+        "name": "processor_config.json",
+        "sha256": "c64dadda906c723a428eedfe6263485a343c35adb64d24295a9b408c372aba31",
+        "sizeBytes": 661
+      },
+      {
+        "name": "tokenizer.json",
+        "sha256": "2ba5b3330fd84d5376fcca797cfb3b42eee6241ce23e3271e6fb2a115a8751bd",
+        "sizeBytes": 17077420
+      },
+      {
+        "name": "tokenizer_config.json",
+        "sha256": "776dd712367ab3da3e79c5008c6141b1f8912c74310410c4ed49827299b856d1",
+        "sizeBytes": 21190
+      }
+    ]
+  },
+  "minGezelVersion": "1.26227"
+}

package/data/chat-models/mu/muse-glimmer-30b-q4/manifest.json

@@ -4,7 +4,7 @@
   "minGezelVersion": "1.26223",
   "id": "muse-glimmer-30b-q4",
   "name": "Muse Glimmer (30B, Q4)",
-  "description": "Meta Superintelligence Labs' Muse family, distilled from the larger Muse Spark down to 30B: a dense agentic model (29.6B language model plus a ~1.8B ViT-G/14 perception encoder) built for long-horizon tool use on consumer hardware. A new architecture rather than a Llama derivative — its own chat format and tool-call grammar. Multimodal (text + image), 131K native context, native tool calling, and a reasoning strength dial from low to xhigh. Meta's own calibrated K-quant build, sized to fit 24GB of VRAM.",
+  "description": "Meta Superintelligence Labs' Muse Glimmer at 30B parameters, distilled from the larger Muse Spark and built for long-horizon tool use on everyday hardware. Multimodal (text + image), 131K context, native tool calling, and adjustable reasoning effort. Meta's own build, sized for machines with 24GB+ memory.",
   "tags": [
     "meta",
     "agentic",
@@ -19,8 +19,8 @@
     "name": "Meta Superintelligence Labs",
     "url": "https://huggingface.co/meta-models/Muse-Glimmer-30B"
   },
-  "version": "1.0.0",
-  "updatedAt": "2026-08-10T00:00:00Z",
+  "version": "1.0.1",
+  "updatedAt": "2026-08-15T00:00:00Z",
   "license": "Apache-2.0",
   "licenseClass": "open",
   "licenseShortName": "Apache 2.0",
@@ -129,7 +129,6 @@
       "filename": "mmproj-kquant.gguf",
       "sha256": "f48b452316f9b213758e8659444029b961a24a07f99a1abb2a9f88b06f7c00c6",
       "sizeBytes": 1400328928
-    },
-    "residentBytes": 21416116224
+    }
   }
 }

package/data/chat-models/mu/muse-glimmer-30b-q4/versions/1.0.1/manifest.json

@@ -0,0 +1,20 @@
+{
+  "schemaVersion": 1,
+  "version": "1.0.1",
+  "releasedAt": "2026-08-15T00:00:00Z",
+  "approxSizeBytes": 16756681056,
+  "llamaCpp": {
+    "huggingfaceRepo": "meta-models/Muse-Glimmer-30B-GGUF",
+    "revision": "93769bc7ab5ad1e9cd22d857e3138cf5d977ae81",
+    "filename": "muse-glimmer-30B-kquant-17gb.gguf",
+    "sha256": "7e9b74b7c8875e9e265695df9613bf6290f2392e479ce740495a129019c488d8",
+    "approxSizeBytes": 16756681056,
+    "quantization": "K-Quant-17GB",
+    "mmproj": {
+      "filename": "mmproj-kquant.gguf",
+      "sha256": "f48b452316f9b213758e8659444029b961a24a07f99a1abb2a9f88b06f7c00c6",
+      "sizeBytes": 1400328928
+    }
+  },
+  "minGezelVersion": "1.26227"
+}

package/data/chat-models/ne/nemotron3-nano-30b-q4/manifest.json

@@ -3,7 +3,7 @@
   "kind": "chat-model",
   "id": "nemotron3-nano-30b-q4",
   "name": "Nemotron-3 Nano (30B, MoE Reasoning)",
-  "description": "NVIDIA's Nemotron-3 Nano Omni — 30B-parameter hybrid Mamba/Transformer Mixture-of-Experts (~3B active per token) tuned for reasoning. Native function-calling, 128K context, multimodal (text + image + audio + video).",
+  "description": "NVIDIA's Nemotron-3 Nano Omni at 30B parameters, tuned for reasoning. Multimodal (text + image + audio + video), 128K context, native tool calling. Capable-tier on-device option for machines with 32GB+ memory.",
   "tags": [
     "nvidia",
     "reasoning",
@@ -20,8 +20,8 @@
     "name": "NVIDIA",
     "url": "https://huggingface.co/nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning"
   },
-  "version": "1.0.0",
-  "updatedAt": "2026-05-17T00:00:00Z",
+  "version": "1.0.2",
+  "updatedAt": "2026-08-15T00:00:00Z",
   "license": "NVIDIA Open Model License",
   "licenseClass": "custom-restricted",
   "licenseShortName": "NVIDIA Open Model",

package/data/chat-models/ne/nemotron3-nano-30b-q4/versions/1.0.1/manifest.json

@@ -0,0 +1,94 @@
+{
+  "schemaVersion": 1,
+  "version": "1.0.1",
+  "releasedAt": "2026-08-15T00:00:00Z",
+  "approxSizeBytes": 24515129536,
+  "llamaCpp": {
+    "huggingfaceRepo": "lmstudio-community/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-GGUF",
+    "revision": "49a5b1ae335e30e775d4eb03f822b52a63ea4271",
+    "filename": "Nemotron-3-Nano-Omni-30B-A3B-Reasoning-Q4_K_M.gguf",
+    "sha256": "e3efd3f173538eef6a422cfb20f273413348adfd75e14b846acf36943894ef43",
+    "approxSizeBytes": 24515129536,
+    "residentBytes": 29418155443,
+    "quantization": "Q4_K_M",
+    "mmproj": {
+      "filename": "mmproj-Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16.gguf",
+      "sha256": "0e46f4126eb65732c645262d78b2ac41955ced3de4643dddfa9892f68d08c9e9",
+      "sizeBytes": 1589506176
+    }
+  },
+  "mlx": {
+    "huggingfaceRepo": "mlx-community/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-nvfp4",
+    "revision": "3f65bc48c603081c6ba89b63fdc23400ea087e38",
+    "quantization": "nvfp4",
+    "approxSizeBytes": 19650740662,
+    "files": [
+      {
+        "name": "chat_template.jinja",
+        "sha256": "41428e0c65e312c359df2495ef5284769a9520b15a693deda4c34a1538208faa",
+        "sizeBytes": 14277
+      },
+      {
+        "name": "config.json",
+        "sha256": "1fb89f29105b9fb455ee8e1dc88de963b95a631c7cc6a68d7e17d4aa24570aac",
+        "sizeBytes": 9459
+      },
+      {
+        "name": "generation_config.json",
+        "sha256": "dbe3dfa851ad1462c45beed3a9a35f02e24af5311d79e8b9de5a41428b3a8e7a",
+        "sizeBytes": 309
+      },
+      {
+        "name": "model-00001-of-00004.safetensors",
+        "sha256": "128a189c3e20c23b5199394b0c84f1dae226405f82081e5d559e3fb2c980b343",
+        "sizeBytes": 5118626596
+      },
+      {
+        "name": "model-00002-of-00004.safetensors",
+        "sha256": "a2097eb987352e665992af87edd101047e713a1f80cc34555254179033f00fe6",
+        "sizeBytes": 5291563052
+      },
+      {
+        "name": "model-00003-of-00004.safetensors",
+        "sha256": "1c77e0cb0d69884a9bb01e540fa43818c10f106bb447edbef5fe7a80b5a4f38b",
+        "sizeBytes": 5291563074
+      },
+      {
+        "name": "model-00004-of-00004.safetensors",
+        "sha256": "903b1ecdef13e3a80fe430a28143ffd4062337c6cd5e78ba9986a127fbd78a25",
+        "sizeBytes": 3931693211
+      },
+      {
+        "name": "model.safetensors.index.json",
+        "sha256": "09a169e3c52670cfe7ea95930139f17abed38144d2d220db5057e2e0ba000f13",
+        "sizeBytes": 191231
+      },
+      {
+        "name": "preprocessor_config.json",
+        "sha256": "e53c54889896ad6104a031f58831cf237e1416bfb16612d856213ce94a91df17",
+        "sizeBytes": 342
+      },
+      {
+        "name": "processor_config.json",
+        "sha256": "f7de807026c387af42011cc85058253b7cf39ddb5cb8cedf8927d7d2eb2c6ce7",
+        "sizeBytes": 743
+      },
+      {
+        "name": "special_tokens_map.json",
+        "sha256": "e3a4f63da745f02317a45e00e6476c17fc66ac41faf14bb1b0be1f3211b0ca53",
+        "sizeBytes": 420
+      },
+      {
+        "name": "tokenizer.json",
+        "sha256": "d0432ad8b1c2984a2e2dca82e1933e3fd5628df6e95c51fba4c3ee9950dce299",
+        "sizeBytes": 17077366
+      },
+      {
+        "name": "tokenizer_config.json",
+        "sha256": "1293fc27661bc2989b8c3b42db6ffbdfad87e35144b9b686eab2538b4c56c60e",
+        "sizeBytes": 582
+      }
+    ]
+  },
+  "minGezelVersion": "1.26227"
+}

package/data/chat-models/ne/nemotron3-nano-30b-q4/versions/1.0.2/manifest.json

@@ -0,0 +1,93 @@
+{
+  "schemaVersion": 1,
+  "version": "1.0.2",
+  "releasedAt": "2026-08-15T00:00:00Z",
+  "approxSizeBytes": 24515129536,
+  "llamaCpp": {
+    "huggingfaceRepo": "lmstudio-community/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-GGUF",
+    "revision": "49a5b1ae335e30e775d4eb03f822b52a63ea4271",
+    "filename": "Nemotron-3-Nano-Omni-30B-A3B-Reasoning-Q4_K_M.gguf",
+    "sha256": "e3efd3f173538eef6a422cfb20f273413348adfd75e14b846acf36943894ef43",
+    "approxSizeBytes": 24515129536,
+    "quantization": "Q4_K_M",
+    "mmproj": {
+      "filename": "mmproj-Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16.gguf",
+      "sha256": "0e46f4126eb65732c645262d78b2ac41955ced3de4643dddfa9892f68d08c9e9",
+      "sizeBytes": 1589506176
+    }
+  },
+  "mlx": {
+    "huggingfaceRepo": "mlx-community/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-nvfp4",
+    "revision": "3f65bc48c603081c6ba89b63fdc23400ea087e38",
+    "quantization": "nvfp4",
+    "approxSizeBytes": 19650740662,
+    "files": [
+      {
+        "name": "chat_template.jinja",
+        "sha256": "41428e0c65e312c359df2495ef5284769a9520b15a693deda4c34a1538208faa",
+        "sizeBytes": 14277
+      },
+      {
+        "name": "config.json",
+        "sha256": "1fb89f29105b9fb455ee8e1dc88de963b95a631c7cc6a68d7e17d4aa24570aac",
+        "sizeBytes": 9459
+      },
+      {
+        "name": "generation_config.json",
+        "sha256": "dbe3dfa851ad1462c45beed3a9a35f02e24af5311d79e8b9de5a41428b3a8e7a",
+        "sizeBytes": 309
+      },
+      {
+        "name": "model-00001-of-00004.safetensors",
+        "sha256": "128a189c3e20c23b5199394b0c84f1dae226405f82081e5d559e3fb2c980b343",
+        "sizeBytes": 5118626596
+      },
+      {
+        "name": "model-00002-of-00004.safetensors",
+        "sha256": "a2097eb987352e665992af87edd101047e713a1f80cc34555254179033f00fe6",
+        "sizeBytes": 5291563052
+      },
+      {
+        "name": "model-00003-of-00004.safetensors",
+        "sha256": "1c77e0cb0d69884a9bb01e540fa43818c10f106bb447edbef5fe7a80b5a4f38b",
+        "sizeBytes": 5291563074
+      },
+      {
+        "name": "model-00004-of-00004.safetensors",
+        "sha256": "903b1ecdef13e3a80fe430a28143ffd4062337c6cd5e78ba9986a127fbd78a25",
+        "sizeBytes": 3931693211
+      },
+      {
+        "name": "model.safetensors.index.json",
+        "sha256": "09a169e3c52670cfe7ea95930139f17abed38144d2d220db5057e2e0ba000f13",
+        "sizeBytes": 191231
+      },
+      {
+        "name": "preprocessor_config.json",
+        "sha256": "e53c54889896ad6104a031f58831cf237e1416bfb16612d856213ce94a91df17",
+        "sizeBytes": 342
+      },
+      {
+        "name": "processor_config.json",
+        "sha256": "f7de807026c387af42011cc85058253b7cf39ddb5cb8cedf8927d7d2eb2c6ce7",
+        "sizeBytes": 743
+      },
+      {
+        "name": "special_tokens_map.json",
+        "sha256": "e3a4f63da745f02317a45e00e6476c17fc66ac41faf14bb1b0be1f3211b0ca53",
+        "sizeBytes": 420
+      },
+      {
+        "name": "tokenizer.json",
+        "sha256": "d0432ad8b1c2984a2e2dca82e1933e3fd5628df6e95c51fba4c3ee9950dce299",
+        "sizeBytes": 17077366
+      },
+      {
+        "name": "tokenizer_config.json",
+        "sha256": "1293fc27661bc2989b8c3b42db6ffbdfad87e35144b9b686eab2538b4c56c60e",
+        "sizeBytes": 582
+      }
+    ]
+  },
+  "minGezelVersion": "1.26227"
+}

package/data/chat-models/ne/nemotron3-super-120b-q4/manifest.json

@@ -3,7 +3,7 @@
   "kind": "chat-model",
   "id": "nemotron3-super-120b-q4",
   "name": "Nemotron-3 Super (120B-A12B)",
-  "description": "NVIDIA's flagship Nemotron-3 Super — 120B-parameter hybrid Mamba/Transformer Mixture-of-Experts (~12B active per token). Native function-calling, 128K context. Frontier-tier on-device reasoning for workstations and Mac Studio-class hardware — needs ~100GB of usable memory for Q4_K_M. MLX path is held until mlx-community publishes a build that doesn't require trust_remote_code; for now this is llama-only.",
+  "description": "NVIDIA's flagship Nemotron-3 Super at 120B parameters, tuned for reasoning. Native tool calling, 128K context. Frontier-tier on-device reasoning for workstations and Mac Studio-class hardware — needs ~100GB of usable memory. No Apple Silicon build yet.",
   "tags": [
     "nvidia",
     "reasoning",
@@ -17,8 +17,8 @@
     "name": "NVIDIA",
     "url": "https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-BF16"
   },
-  "version": "1.0.0",
-  "updatedAt": "2026-05-17T00:00:00Z",
+  "version": "1.0.1",
+  "updatedAt": "2026-08-15T00:00:00Z",
   "license": "NVIDIA Open Model License",
   "licenseClass": "custom-restricted",
   "licenseShortName": "NVIDIA Open Model",

package/data/chat-models/ne/nemotron3-super-120b-q4/versions/1.0.1/manifest.json

@@ -0,0 +1,30 @@
+{
+  "schemaVersion": 1,
+  "version": "1.0.1",
+  "releasedAt": "2026-08-15T00:00:00Z",
+  "approxSizeBytes": 86051079584,
+  "llamaCpp": {
+    "huggingfaceRepo": "lmstudio-community/NVIDIA-Nemotron-3-Super-120B-A12B-GGUF",
+    "revision": "16410500b9bdc62081c34cc2dff52471e001f970",
+    "shards": [
+      {
+        "name": "NVIDIA-Nemotron-3-Super-120B-A12B-Q4_K_M-00001-of-00003.gguf",
+        "sha256": "2cf8e99b9b71babc5932eb594b1fe1f9b9e407a460a5b49b0625f0c25e0ba799",
+        "sizeBytes": 39205633408
+      },
+      {
+        "name": "NVIDIA-Nemotron-3-Super-120B-A12B-Q4_K_M-00002-of-00003.gguf",
+        "sha256": "770b505e6fe3b19b422b5b0dd06201b07218d0af99e72e69086208983fe54219",
+        "sizeBytes": 39720681760
+      },
+      {
+        "name": "NVIDIA-Nemotron-3-Super-120B-A12B-Q4_K_M-00003-of-00003.gguf",
+        "sha256": "7420107ccfc77230deaf57971ae84a58a18c4e6e035aa99cf0e25fbbc4c7614a",
+        "sizeBytes": 7124764416
+      }
+    ],
+    "approxSizeBytes": 86051079584,
+    "quantization": "Q4_K_M"
+  },
+  "minGezelVersion": "1.26227"
+}

package/data/chat-models/ne/nemotron3.5-lightning-30b-q4/manifest.json

@@ -3,7 +3,7 @@
   "kind": "chat-model",
   "id": "nemotron3.5-lightning-30b-q4",
   "name": "Nemotron 3.5 Lightning (30B-A3B, Q4)",
-  "description": "NVIDIA's Nemotron 3.5 Lightning — a text-only 30B-parameter hybrid Mamba-2/Transformer Mixture-of-Experts model with about 3B active parameters per token, native reasoning and tool calling, and a 256K context window. The llama.cpp source is ggml-org's Q4_K_M build (about 25.4GB); Apple Silicon uses mlx-community's matching 4-bit affine conversion (about 17.8GB).",
+  "description": "NVIDIA's Nemotron 3.5 Lightning at 30B parameters — text-only, with native reasoning, native tool calling, and a 256K context window. Capable-tier on-device option: about 25GB of memory on most machines, about 18GB on Apple Silicon.",
   "tags": [
     "nvidia",
     "reasoning",
@@ -22,8 +22,8 @@
     "name": "NVIDIA",
     "url": "https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16"
   },
-  "version": "1.0.0",
-  "updatedAt": "2026-08-11T00:00:00Z",
+  "version": "1.0.1",
+  "updatedAt": "2026-08-15T00:00:00Z",
   "minGezelVersion": "1.26223",
   "license": "OpenMDW-1.1",
   "licenseClass": "open",
@@ -129,8 +129,7 @@
     "filename": "NVIDIA-Nemotron-3.5-Lightning-30B-A3B-Q4_K_M.gguf",
     "sha256": "6110e2e2e6cd324e6ee69ddced5a6b34fad6c94ca9827222a1e420fb92e3c90b",
     "approxSizeBytes": 25430738944,
-    "quantization": "Q4_K_M",
-    "residentBytes": 30516886733
+    "quantization": "Q4_K_M"
   },
   "mlx": {
     "huggingfaceRepo": "mlx-community/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-4bit",
@@ -188,7 +187,6 @@
         "sha256": "ec380f0a6c0c2e294eb2435f5ec521fb2121cfc970e6ebf023d06829714a4eaa",
         "sizeBytes": 430
       }
-    ],
-    "residentBytes": 23130356200
+    ]
   }
 }