Switch catalog to per-source entries

Replace the single catalog_v1.json snapshot with one entry file per source repo under entries/<owner>/<repo>.json.

entries/Qwen/Qwen2.5-0.5B-Instruct-GGUF.json

@@ -0,0 +1,23 @@
+{
+  "schema_version": 1,
+  "source_repo": "Qwen/Qwen2.5-0.5B-Instruct-GGUF",
+  "variants": {
+    "qwen2.5-0.5b-instruct-q4_k_m": {
+      "source": {
+        "repo": "Qwen/Qwen2.5-0.5B-Instruct-GGUF",
+        "revision": "main",
+        "file": "qwen2.5-0.5b-instruct-q4_k_m.gguf"
+      },
+      "curated": {
+        "name": "Qwen2.5-0.5B-Instruct-Q4_K_M",
+        "size": "491MB",
+        "description": "Draft for Qwen2.5 and DeepSeek-R1-Distill models",
+        "draft": null,
+        "moe": null,
+        "extra_files": [],
+        "mmproj": null
+      },
+      "packages": []
+    }
+  }
+}

entries/Qwen/Qwen2.5-3B-Instruct-GGUF.json

@@ -0,0 +1,23 @@
+{
+  "schema_version": 1,
+  "source_repo": "Qwen/Qwen2.5-3B-Instruct-GGUF",
+  "variants": {
+    "qwen2.5-3b-instruct-q4_k_m": {
+      "source": {
+        "repo": "Qwen/Qwen2.5-3B-Instruct-GGUF",
+        "revision": "main",
+        "file": "qwen2.5-3b-instruct-q4_k_m.gguf"
+      },
+      "curated": {
+        "name": "Qwen2.5-3B-Instruct-Q4_K_M",
+        "size": "2.1GB",
+        "description": "Small & fast general chat",
+        "draft": "Qwen2.5-0.5B-Instruct-Q4_K_M",
+        "moe": null,
+        "extra_files": [],
+        "mmproj": null
+      },
+      "packages": []
+    }
+  }
+}

entries/Qwen/Qwen2.5-Coder-32B-Instruct-GGUF.json

@@ -0,0 +1,23 @@
+{
+  "schema_version": 1,
+  "source_repo": "Qwen/Qwen2.5-Coder-32B-Instruct-GGUF",
+  "variants": {
+    "qwen2.5-coder-32b-instruct-q4_k_m": {
+      "source": {
+        "repo": "Qwen/Qwen2.5-Coder-32B-Instruct-GGUF",
+        "revision": "main",
+        "file": "qwen2.5-coder-32b-instruct-q4_k_m.gguf"
+      },
+      "curated": {
+        "name": "Qwen2.5-Coder-32B-Instruct-Q4_K_M",
+        "size": "20GB",
+        "description": "Top-tier code gen, matches GPT-4o on code",
+        "draft": "Qwen2.5-0.5B-Instruct-Q4_K_M",
+        "moe": null,
+        "extra_files": [],
+        "mmproj": null
+      },
+      "packages": []
+    }
+  }
+}

entries/Qwen/Qwen2.5-Coder-7B-Instruct-GGUF.json

@@ -0,0 +1,23 @@
+{
+  "schema_version": 1,
+  "source_repo": "Qwen/Qwen2.5-Coder-7B-Instruct-GGUF",
+  "variants": {
+    "qwen2.5-coder-7b-instruct-q4_k_m": {
+      "source": {
+        "repo": "Qwen/Qwen2.5-Coder-7B-Instruct-GGUF",
+        "revision": "main",
+        "file": "qwen2.5-coder-7b-instruct-q4_k_m.gguf"
+      },
+      "curated": {
+        "name": "Qwen2.5-Coder-7B-Instruct-Q4_K_M",
+        "size": "4.4GB",
+        "description": "Code generation & completion",
+        "draft": "Qwen2.5-0.5B-Instruct-Q4_K_M",
+        "moe": null,
+        "extra_files": [],
+        "mmproj": null
+      },
+      "packages": []
+    }
+  }
+}

entries/Qwen/Qwen3-Coder-Next-GGUF.json

@@ -0,0 +1,39 @@
+{
+  "schema_version": 1,
+  "source_repo": "Qwen/Qwen3-Coder-Next-GGUF",
+  "variants": {
+    "Qwen3-Coder-Next-Q4_K_M/Qwen3-Coder-Next-Q4_K_M-00001-of-00004": {
+      "source": {
+        "repo": "Qwen/Qwen3-Coder-Next-GGUF",
+        "revision": "main",
+        "file": "Qwen3-Coder-Next-Q4_K_M/Qwen3-Coder-Next-Q4_K_M-00001-of-00004.gguf"
+      },
+      "curated": {
+        "name": "Qwen3-Coder-Next-Q4_K_M",
+        "size": "48GB",
+        "description": "Qwen3 Coder Next ~85B dense, frontier coding model",
+        "draft": null,
+        "moe": null,
+        "extra_files": [
+          {
+            "file": "Qwen3-Coder-Next-Q4_K_M-00002-of-00004.gguf",
+            "repo": "Qwen/Qwen3-Coder-Next-GGUF",
+            "revision": "main"
+          },
+          {
+            "file": "Qwen3-Coder-Next-Q4_K_M-00003-of-00004.gguf",
+            "repo": "Qwen/Qwen3-Coder-Next-GGUF",
+            "revision": "main"
+          },
+          {
+            "file": "Qwen3-Coder-Next-Q4_K_M-00004-of-00004.gguf",
+            "repo": "Qwen/Qwen3-Coder-Next-GGUF",
+            "revision": "main"
+          }
+        ],
+        "mmproj": null
+      },
+      "packages": []
+    }
+  }
+}

entries/bartowski/Hermes-2-Pro-Mistral-7B-GGUF.json

@@ -0,0 +1,23 @@
+{
+  "schema_version": 1,
+  "source_repo": "bartowski/Hermes-2-Pro-Mistral-7B-GGUF",
+  "variants": {
+    "Hermes-2-Pro-Mistral-7B-Q4_K_M": {
+      "source": {
+        "repo": "bartowski/Hermes-2-Pro-Mistral-7B-GGUF",
+        "revision": "main",
+        "file": "Hermes-2-Pro-Mistral-7B-Q4_K_M.gguf"
+      },
+      "curated": {
+        "name": "Hermes-2-Pro-Mistral-7B-Q4_K_M",
+        "size": "4.4GB",
+        "description": "Goose default, strong tool calling for agents",
+        "draft": null,
+        "moe": null,
+        "extra_files": [],
+        "mmproj": null
+      },
+      "packages": []
+    }
+  }
+}

entries/bartowski/Llama-3.2-1B-Instruct-GGUF.json

@@ -0,0 +1,23 @@
+{
+  "schema_version": 1,
+  "source_repo": "bartowski/Llama-3.2-1B-Instruct-GGUF",
+  "variants": {
+    "Llama-3.2-1B-Instruct-Q4_K_M": {
+      "source": {
+        "repo": "bartowski/Llama-3.2-1B-Instruct-GGUF",
+        "revision": "main",
+        "file": "Llama-3.2-1B-Instruct-Q4_K_M.gguf"
+      },
+      "curated": {
+        "name": "Llama-3.2-1B-Instruct-Q4_K_M",
+        "size": "760MB",
+        "description": "Draft for Llama 3.x and Llama 4 models",
+        "draft": null,
+        "moe": null,
+        "extra_files": [],
+        "mmproj": null
+      },
+      "packages": []
+    }
+  }
+}

entries/bartowski/Llama-3.2-3B-Instruct-GGUF.json

@@ -0,0 +1,23 @@
+{
+  "schema_version": 1,
+  "source_repo": "bartowski/Llama-3.2-3B-Instruct-GGUF",
+  "variants": {
+    "Llama-3.2-3B-Instruct-Q4_K_M": {
+      "source": {
+        "repo": "bartowski/Llama-3.2-3B-Instruct-GGUF",
+        "revision": "main",
+        "file": "Llama-3.2-3B-Instruct-Q4_K_M.gguf"
+      },
+      "curated": {
+        "name": "Llama-3.2-3B-Instruct-Q4_K_M",
+        "size": "2.0GB",
+        "description": "Meta Llama 3.2, goose default, good tool calling",
+        "draft": "Llama-3.2-1B-Instruct-Q4_K_M",
+        "moe": null,
+        "extra_files": [],
+        "mmproj": null
+      },
+      "packages": []
+    }
+  }
+}

entries/bartowski/Qwen2.5-14B-Instruct-GGUF.json

@@ -0,0 +1,23 @@
+{
+  "schema_version": 1,
+  "source_repo": "bartowski/Qwen2.5-14B-Instruct-GGUF",
+  "variants": {
+    "Qwen2.5-14B-Instruct-Q4_K_M": {
+      "source": {
+        "repo": "bartowski/Qwen2.5-14B-Instruct-GGUF",
+        "revision": "main",
+        "file": "Qwen2.5-14B-Instruct-Q4_K_M.gguf"
+      },
+      "curated": {
+        "name": "Qwen2.5-14B-Instruct-Q4_K_M",
+        "size": "9.0GB",
+        "description": "Solid general chat",
+        "draft": "Qwen2.5-0.5B-Instruct-Q4_K_M",
+        "moe": null,
+        "extra_files": [],
+        "mmproj": null
+      },
+      "packages": []
+    }
+  }
+}

entries/bartowski/Qwen2.5-32B-Instruct-GGUF.json

@@ -0,0 +1,23 @@
+{
+  "schema_version": 1,
+  "source_repo": "bartowski/Qwen2.5-32B-Instruct-GGUF",
+  "variants": {
+    "Qwen2.5-32B-Instruct-Q4_K_M": {
+      "source": {
+        "repo": "bartowski/Qwen2.5-32B-Instruct-GGUF",
+        "revision": "main",
+        "file": "Qwen2.5-32B-Instruct-Q4_K_M.gguf"
+      },
+      "curated": {
+        "name": "Qwen2.5-32B-Instruct-Q4_K_M",
+        "size": "20GB",
+        "description": "Proven general chat",
+        "draft": "Qwen2.5-0.5B-Instruct-Q4_K_M",
+        "moe": null,
+        "extra_files": [],
+        "mmproj": null
+      },
+      "packages": []
+    }
+  }
+}

entries/bartowski/Qwen2.5-Coder-14B-Instruct-GGUF.json

@@ -0,0 +1,23 @@
+{
+  "schema_version": 1,
+  "source_repo": "bartowski/Qwen2.5-Coder-14B-Instruct-GGUF",
+  "variants": {
+    "Qwen2.5-Coder-14B-Instruct-Q4_K_M": {
+      "source": {
+        "repo": "bartowski/Qwen2.5-Coder-14B-Instruct-GGUF",
+        "revision": "main",
+        "file": "Qwen2.5-Coder-14B-Instruct-Q4_K_M.gguf"
+      },
+      "curated": {
+        "name": "Qwen2.5-Coder-14B-Instruct-Q4_K_M",
+        "size": "9.0GB",
+        "description": "Strong code gen, fills gap between 7B and 32B",
+        "draft": "Qwen2.5-0.5B-Instruct-Q4_K_M",
+        "moe": null,
+        "extra_files": [],
+        "mmproj": null
+      },
+      "packages": []
+    }
+  }
+}

entries/bartowski/google_gemma-3-1b-it-GGUF.json

@@ -0,0 +1,23 @@
+{
+  "schema_version": 1,
+  "source_repo": "bartowski/google_gemma-3-1b-it-GGUF",
+  "variants": {
+    "google_gemma-3-1b-it-Q4_K_M": {
+      "source": {
+        "repo": "bartowski/google_gemma-3-1b-it-GGUF",
+        "revision": "main",
+        "file": "google_gemma-3-1b-it-Q4_K_M.gguf"
+      },
+      "curated": {
+        "name": "Gemma-3-1B-it-Q4_K_M",
+        "size": "780MB",
+        "description": "Draft for Gemma 3 models",
+        "draft": null,
+        "moe": null,
+        "extra_files": [],
+        "mmproj": null
+      },
+      "packages": []
+    }
+  }
+}

entries/bartowski/google_gemma-3-27b-it-GGUF.json

@@ -0,0 +1,23 @@
+{
+  "schema_version": 1,
+  "source_repo": "bartowski/google_gemma-3-27b-it-GGUF",
+  "variants": {
+    "google_gemma-3-27b-it-Q4_K_M": {
+      "source": {
+        "repo": "bartowski/google_gemma-3-27b-it-GGUF",
+        "revision": "main",
+        "file": "google_gemma-3-27b-it-Q4_K_M.gguf"
+      },
+      "curated": {
+        "name": "Gemma-3-27B-it-Q4_K_M",
+        "size": "17GB",
+        "description": "Google Gemma 3 27B, strong reasoning",
+        "draft": "Gemma-3-1B-it-Q4_K_M",
+        "moe": null,
+        "extra_files": [],
+        "mmproj": null
+      },
+      "packages": []
+    }
+  }
+}

entries/glogwa68/Llama-4-scout-GGUF.json

@@ -0,0 +1,27 @@
+{
+  "schema_version": 1,
+  "source_repo": "glogwa68/Llama-4-scout-GGUF",
+  "variants": {
+    "Llama-4-Scout-4bit-Q4_K_M": {
+      "source": {
+        "repo": "glogwa68/Llama-4-scout-GGUF",
+        "revision": "main",
+        "file": "Llama-4-Scout-4bit-Q4_K_M.gguf"
+      },
+      "curated": {
+        "name": "Llama-4-Scout-Q4_K_M",
+        "size": "22.5GB",
+        "description": "MoE 109B/17B active, 16 experts top-1, Meta latest, tool calling",
+        "draft": null,
+        "moe": {
+          "n_expert": 16,
+          "n_expert_used": 1,
+          "min_experts_per_node": 6
+        },
+        "extra_files": [],
+        "mmproj": null
+      },
+      "packages": []
+    }
+  }
+}

entries/unsloth/DeepSeek-R1-Distill-Qwen-14B-GGUF.json

@@ -0,0 +1,23 @@
+{
+  "schema_version": 1,
+  "source_repo": "unsloth/DeepSeek-R1-Distill-Qwen-14B-GGUF",
+  "variants": {
+    "DeepSeek-R1-Distill-Qwen-14B-Q4_K_M": {
+      "source": {
+        "repo": "unsloth/DeepSeek-R1-Distill-Qwen-14B-GGUF",
+        "revision": "main",
+        "file": "DeepSeek-R1-Distill-Qwen-14B-Q4_K_M.gguf"
+      },
+      "curated": {
+        "name": "DeepSeek-R1-Distill-Qwen-14B-Q4_K_M",
+        "size": "9.0GB",
+        "description": "DeepSeek R1 reasoning distilled into Qwen 14B",
+        "draft": "Qwen2.5-0.5B-Instruct-Q4_K_M",
+        "moe": null,
+        "extra_files": [],
+        "mmproj": null
+      },
+      "packages": []
+    }
+  }
+}

entries/unsloth/DeepSeek-R1-Distill-Qwen-32B-GGUF.json

@@ -0,0 +1,23 @@
+{
+  "schema_version": 1,
+  "source_repo": "unsloth/DeepSeek-R1-Distill-Qwen-32B-GGUF",
+  "variants": {
+    "DeepSeek-R1-Distill-Qwen-32B-Q4_K_M": {
+      "source": {
+        "repo": "unsloth/DeepSeek-R1-Distill-Qwen-32B-GGUF",
+        "revision": "main",
+        "file": "DeepSeek-R1-Distill-Qwen-32B-Q4_K_M.gguf"
+      },
+      "curated": {
+        "name": "DeepSeek-R1-Distill-Qwen-32B-Q4_K_M",
+        "size": "19.9GB",
+        "description": "DeepSeek R1 reasoning distilled into Qwen 32B",
+        "draft": "Qwen2.5-0.5B-Instruct-Q4_K_M",
+        "moe": null,
+        "extra_files": [],
+        "mmproj": null
+      },
+      "packages": []
+    }
+  }
+}

entries/unsloth/Devstral-Small-2505-GGUF.json

@@ -0,0 +1,23 @@
+{
+  "schema_version": 1,
+  "source_repo": "unsloth/Devstral-Small-2505-GGUF",
+  "variants": {
+    "Devstral-Small-2505-Q4_K_M": {
+      "source": {
+        "repo": "unsloth/Devstral-Small-2505-GGUF",
+        "revision": "main",
+        "file": "Devstral-Small-2505-Q4_K_M.gguf"
+      },
+      "curated": {
+        "name": "Devstral-Small-2505-Q4_K_M",
+        "size": "14.3GB",
+        "description": "Mistral agentic coding, tool use",
+        "draft": null,
+        "moe": null,
+        "extra_files": [],
+        "mmproj": null
+      },
+      "packages": []
+    }
+  }
+}

entries/unsloth/GLM-4-32B-0414-GGUF.json

@@ -0,0 +1,23 @@
+{
+  "schema_version": 1,
+  "source_repo": "unsloth/GLM-4-32B-0414-GGUF",
+  "variants": {
+    "GLM-4-32B-0414-Q4_K_M": {
+      "source": {
+        "repo": "unsloth/GLM-4-32B-0414-GGUF",
+        "revision": "main",
+        "file": "GLM-4-32B-0414-Q4_K_M.gguf"
+      },
+      "curated": {
+        "name": "GLM-4-32B-0414-Q4_K_M",
+        "size": "19.7GB",
+        "description": "Strong 32B, good tool calling",
+        "draft": null,
+        "moe": null,
+        "extra_files": [],
+        "mmproj": null
+      },
+      "packages": []
+    }
+  }
+}

entries/unsloth/GLM-4.7-Flash-GGUF.json

@@ -0,0 +1,27 @@
+{
+  "schema_version": 1,
+  "source_repo": "unsloth/GLM-4.7-Flash-GGUF",
+  "variants": {
+    "GLM-4.7-Flash-Q4_K_M": {
+      "source": {
+        "repo": "unsloth/GLM-4.7-Flash-GGUF",
+        "revision": "main",
+        "file": "GLM-4.7-Flash-Q4_K_M.gguf"
+      },
+      "curated": {
+        "name": "GLM-4.7-Flash-Q4_K_M",
+        "size": "18GB",
+        "description": "MoE 30B/3B active, 64 experts top-4, fast inference, tool calling",
+        "draft": null,
+        "moe": {
+          "n_expert": 64,
+          "n_expert_used": 4,
+          "min_experts_per_node": 24
+        },
+        "extra_files": [],
+        "mmproj": null
+      },
+      "packages": []
+    }
+  }
+}

entries/unsloth/MiniMax-M2.5-GGUF.json

@@ -0,0 +1,43 @@
+{
+  "schema_version": 1,
+  "source_repo": "unsloth/MiniMax-M2.5-GGUF",
+  "variants": {
+    "Q4_K_M/MiniMax-M2.5-Q4_K_M-00001-of-00004": {
+      "source": {
+        "repo": "unsloth/MiniMax-M2.5-GGUF",
+        "revision": "main",
+        "file": "Q4_K_M/MiniMax-M2.5-Q4_K_M-00001-of-00004.gguf"
+      },
+      "curated": {
+        "name": "MiniMax-M2.5-Q4_K_M",
+        "size": "138GB",
+        "description": "MiniMax-M2.5 MoE 456B/46B active, 256 experts top-8, Q4_K_M",
+        "draft": null,
+        "moe": {
+          "n_expert": 256,
+          "n_expert_used": 8,
+          "min_experts_per_node": 96
+        },
+        "extra_files": [
+          {
+            "file": "MiniMax-M2.5-Q4_K_M-00002-of-00004.gguf",
+            "repo": "unsloth/MiniMax-M2.5-GGUF",
+            "revision": "main"
+          },
+          {
+            "file": "MiniMax-M2.5-Q4_K_M-00003-of-00004.gguf",
+            "repo": "unsloth/MiniMax-M2.5-GGUF",
+            "revision": "main"
+          },
+          {
+            "file": "MiniMax-M2.5-Q4_K_M-00004-of-00004.gguf",
+            "repo": "unsloth/MiniMax-M2.5-GGUF",
+            "revision": "main"
+          }
+        ],
+        "mmproj": null
+      },
+      "packages": []
+    }
+  }
+}

entries/unsloth/Mistral-Small-3.1-24B-Instruct-2503-GGUF.json

@@ -0,0 +1,23 @@
+{
+  "schema_version": 1,
+  "source_repo": "unsloth/Mistral-Small-3.1-24B-Instruct-2503-GGUF",
+  "variants": {
+    "Mistral-Small-3.1-24B-Instruct-2503-Q4_K_M": {
+      "source": {
+        "repo": "unsloth/Mistral-Small-3.1-24B-Instruct-2503-GGUF",
+        "revision": "main",
+        "file": "Mistral-Small-3.1-24B-Instruct-2503-Q4_K_M.gguf"
+      },
+      "curated": {
+        "name": "Mistral-Small-3.1-24B-Instruct-Q4_K_M",
+        "size": "14.3GB",
+        "description": "Mistral general chat, good tool calling",
+        "draft": null,
+        "moe": null,
+        "extra_files": [],
+        "mmproj": null
+      },
+      "packages": []
+    }
+  }
+}

entries/unsloth/Qwen3-0.6B-GGUF.json

@@ -0,0 +1,23 @@
+{
+  "schema_version": 1,
+  "source_repo": "unsloth/Qwen3-0.6B-GGUF",
+  "variants": {
+    "Qwen3-0.6B-Q4_K_M": {
+      "source": {
+        "repo": "unsloth/Qwen3-0.6B-GGUF",
+        "revision": "main",
+        "file": "Qwen3-0.6B-Q4_K_M.gguf"
+      },
+      "curated": {
+        "name": "Qwen3-0.6B-Q4_K_M",
+        "size": "397MB",
+        "description": "Draft for Qwen3 models",
+        "draft": null,
+        "moe": null,
+        "extra_files": [],
+        "mmproj": null
+      },
+      "packages": []
+    }
+  }
+}

entries/unsloth/Qwen3-14B-GGUF.json

@@ -0,0 +1,23 @@
+{
+  "schema_version": 1,
+  "source_repo": "unsloth/Qwen3-14B-GGUF",
+  "variants": {
+    "Qwen3-14B-Q4_K_M": {
+      "source": {
+        "repo": "unsloth/Qwen3-14B-GGUF",
+        "revision": "main",
+        "file": "Qwen3-14B-Q4_K_M.gguf"
+      },
+      "curated": {
+        "name": "Qwen3-14B-Q4_K_M",
+        "size": "9.0GB",
+        "description": "Qwen3 strong chat, thinking modes",
+        "draft": "Qwen3-0.6B-Q4_K_M",
+        "moe": null,
+        "extra_files": [],
+        "mmproj": null
+      },
+      "packages": []
+    }
+  }
+}

entries/unsloth/Qwen3-30B-A3B-GGUF.json

@@ -0,0 +1,27 @@
+{
+  "schema_version": 1,
+  "source_repo": "unsloth/Qwen3-30B-A3B-GGUF",
+  "variants": {
+    "Qwen3-30B-A3B-Q4_K_M": {
+      "source": {
+        "repo": "unsloth/Qwen3-30B-A3B-GGUF",
+        "revision": "main",
+        "file": "Qwen3-30B-A3B-Q4_K_M.gguf"
+      },
+      "curated": {
+        "name": "Qwen3-30B-A3B-Q4_K_M",
+        "size": "17.3GB",
+        "description": "MoE general chat, 128 experts top-8, thinking/non-thinking",
+        "draft": "Qwen3-0.6B-Q4_K_M",
+        "moe": {
+          "n_expert": 128,
+          "n_expert_used": 8,
+          "min_experts_per_node": 46
+        },
+        "extra_files": [],
+        "mmproj": null
+      },
+      "packages": []
+    }
+  }
+}

entries/unsloth/Qwen3-32B-GGUF.json

@@ -0,0 +1,23 @@
+{
+  "schema_version": 1,
+  "source_repo": "unsloth/Qwen3-32B-GGUF",
+  "variants": {
+    "Qwen3-32B-Q4_K_M": {
+      "source": {
+        "repo": "unsloth/Qwen3-32B-GGUF",
+        "revision": "main",
+        "file": "Qwen3-32B-Q4_K_M.gguf"
+      },
+      "curated": {
+        "name": "Qwen3-32B-Q4_K_M",
+        "size": "19.8GB",
+        "description": "Best Qwen3 dense, thinking/non-thinking modes",
+        "draft": "Qwen3-0.6B-Q4_K_M",
+        "moe": null,
+        "extra_files": [],
+        "mmproj": null
+      },
+      "packages": []
+    }
+  }
+}

entries/unsloth/Qwen3-4B-GGUF.json

@@ -0,0 +1,23 @@
+{
+  "schema_version": 1,
+  "source_repo": "unsloth/Qwen3-4B-GGUF",
+  "variants": {
+    "Qwen3-4B-Q4_K_M": {
+      "source": {
+        "repo": "unsloth/Qwen3-4B-GGUF",
+        "revision": "main",
+        "file": "Qwen3-4B-Q4_K_M.gguf"
+      },
+      "curated": {
+        "name": "Qwen3-4B-Q4_K_M",
+        "size": "2.5GB",
+        "description": "Qwen3 starter, thinking/non-thinking modes",
+        "draft": "Qwen3-0.6B-Q4_K_M",
+        "moe": null,
+        "extra_files": [],
+        "mmproj": null
+      },
+      "packages": []
+    }
+  }
+}

entries/unsloth/Qwen3-8B-GGUF.json

@@ -0,0 +1,23 @@
+{
+  "schema_version": 1,
+  "source_repo": "unsloth/Qwen3-8B-GGUF",
+  "variants": {
+    "Qwen3-8B-Q4_K_M": {
+      "source": {
+        "repo": "unsloth/Qwen3-8B-GGUF",
+        "revision": "main",
+        "file": "Qwen3-8B-Q4_K_M.gguf"
+      },
+      "curated": {
+        "name": "Qwen3-8B-Q4_K_M",
+        "size": "5.0GB",
+        "description": "Qwen3 mid-tier, strong for its size",
+        "draft": "Qwen3-0.6B-Q4_K_M",
+        "moe": null,
+        "extra_files": [],
+        "mmproj": null
+      },
+      "packages": []
+    }
+  }
+}

entries/unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF.json

@@ -0,0 +1,27 @@
+{
+  "schema_version": 1,
+  "source_repo": "unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF",
+  "variants": {
+    "Qwen3-Coder-30B-A3B-Instruct-Q4_K_M": {
+      "source": {
+        "repo": "unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF",
+        "revision": "main",
+        "file": "Qwen3-Coder-30B-A3B-Instruct-Q4_K_M.gguf"
+      },
+      "curated": {
+        "name": "Qwen3-Coder-30B-A3B-Instruct-Q4_K_M",
+        "size": "18.6GB",
+        "description": "MoE agentic coding, tool use, 128 experts top-8",
+        "draft": "Qwen3-0.6B-Q4_K_M",
+        "moe": {
+          "n_expert": 128,
+          "n_expert_used": 8,
+          "min_experts_per_node": 46
+        },
+        "extra_files": [],
+        "mmproj": null
+      },
+      "packages": []
+    }
+  }
+}

entries/unsloth/Qwen3.5-0.8B-GGUF.json

@@ -0,0 +1,27 @@
+{
+  "schema_version": 1,
+  "source_repo": "unsloth/Qwen3.5-0.8B-GGUF",
+  "variants": {
+    "Qwen3.5-0.8B-Q4_K_M": {
+      "source": {
+        "repo": "unsloth/Qwen3.5-0.8B-GGUF",
+        "revision": "main",
+        "file": "Qwen3.5-0.8B-Q4_K_M.gguf"
+      },
+      "curated": {
+        "name": "Qwen3.5-0.8B-Vision-Q4_K_M",
+        "size": "508MB",
+        "description": "Tiny vision model, OCR, screenshots, runs anywhere",
+        "draft": null,
+        "moe": null,
+        "extra_files": [],
+        "mmproj": {
+          "file": "Qwen3.5-0.8B-mmproj-BF16.gguf",
+          "repo": "unsloth/Qwen3.5-0.8B-GGUF",
+          "revision": "main"
+        }
+      },
+      "packages": []
+    }
+  }
+}

entries/unsloth/Qwen3.5-4B-GGUF.json

@@ -0,0 +1,27 @@
+{
+  "schema_version": 1,
+  "source_repo": "unsloth/Qwen3.5-4B-GGUF",
+  "variants": {
+    "Qwen3.5-4B-Q4_K_M": {
+      "source": {
+        "repo": "unsloth/Qwen3.5-4B-GGUF",
+        "revision": "main",
+        "file": "Qwen3.5-4B-Q4_K_M.gguf"
+      },
+      "curated": {
+        "name": "Qwen3.5-4B-Vision-Q4_K_M",
+        "size": "2.7GB",
+        "description": "Small vision model, good quality/size balance",
+        "draft": null,
+        "moe": null,
+        "extra_files": [],
+        "mmproj": {
+          "file": "Qwen3.5-4B-mmproj-BF16.gguf",
+          "repo": "unsloth/Qwen3.5-4B-GGUF",
+          "revision": "main"
+        }
+      },
+      "packages": []
+    }
+  }
+}

entries/unsloth/Qwen3.5-9B-GGUF.json

@@ -0,0 +1,27 @@
+{
+  "schema_version": 1,
+  "source_repo": "unsloth/Qwen3.5-9B-GGUF",
+  "variants": {
+    "Qwen3.5-9B-Q4_K_M": {
+      "source": {
+        "repo": "unsloth/Qwen3.5-9B-GGUF",
+        "revision": "main",
+        "file": "Qwen3.5-9B-Q4_K_M.gguf"
+      },
+      "curated": {
+        "name": "Qwen3.5-9B-Vision-Q4_K_M",
+        "size": "5.8GB",
+        "description": "Vision + text, replaces Qwen3-8B with image understanding",
+        "draft": null,
+        "moe": null,
+        "extra_files": [],
+        "mmproj": {
+          "file": "Qwen3.5-9B-mmproj-BF16.gguf",
+          "repo": "unsloth/Qwen3.5-9B-GGUF",
+          "revision": "main"
+        }
+      },
+      "packages": []
+    }
+  }
+}

entries/unsloth/gemma-3-12b-it-GGUF.json

@@ -0,0 +1,23 @@
+{
+  "schema_version": 1,
+  "source_repo": "unsloth/gemma-3-12b-it-GGUF",
+  "variants": {
+    "gemma-3-12b-it-Q4_K_M": {
+      "source": {
+        "repo": "unsloth/gemma-3-12b-it-GGUF",
+        "revision": "main",
+        "file": "gemma-3-12b-it-Q4_K_M.gguf"
+      },
+      "curated": {
+        "name": "Gemma-3-12B-it-Q4_K_M",
+        "size": "7.3GB",
+        "description": "Google Gemma 3 12B, punches above weight",
+        "draft": "Gemma-3-1B-it-Q4_K_M",
+        "moe": null,
+        "extra_files": [],
+        "mmproj": null
+      },
+      "packages": []
+    }
+  }
+}