{
"catalog_version": 2,
"updated": "2026-07-10T00:00:00Z",
"entries": [
{
"name": "qwen3-0.6b",
"task": "chat",
"repo": "Qwen/Qwen3-0.6B-GGUF",
"filename": "Qwen3-0.6B-Q8_0.gguf",
"quant": "Q8_0",
"file_bytes": 639446688,
"params_b": 0.6,
"japanese": true,
"notes": "Tiny fallback for very low-memory machines."
},
{
"name": "qwen3-1.7b",
"task": "chat",
"repo": "Qwen/Qwen3-1.7B-GGUF",
"filename": "Qwen3-1.7B-Q8_0.gguf",
"quant": "Q8_0",
"file_bytes": 1834426016,
"params_b": 1.7,
"japanese": true,
"notes": "Light general chat."
},
{
"name": "qwen3-4b",
"task": "chat",
"repo": "Qwen/Qwen3-4B-GGUF",
"filename": "Qwen3-4B-Q4_K_M.gguf",
"quant": "Q4_K_M",
"file_bytes": 2497280256,
"params_b": 4.0,
"japanese": true,
"notes": "Good small all-rounder."
},
{
"name": "qwen3-8b",
"task": "chat",
"repo": "Qwen/Qwen3-8B-GGUF",
"filename": "Qwen3-8B-Q4_K_M.gguf",
"quant": "Q4_K_M",
"file_bytes": 5027783488,
"params_b": 8.0,
"japanese": true,
"notes": "Strong general chat."
},
{
"name": "gemma-4-e4b",
"task": "chat",
"repo": "google/gemma-4-E4B-it-qat-q4_0-gguf",
"filename": "gemma-4-E4B_q4_0-it.gguf",
"quant": "Q4_0",
"file_bytes": 5154939136,
"params_b": 8.0,
"japanese": true,
"notes": "Gemma 4 MatFormer (effective 4B), quantization-aware trained."
},
{
"name": "gemma-4-12b",
"task": "chat",
"repo": "google/gemma-4-12B-it-qat-q4_0-gguf",
"filename": "gemma-4-12b-it-qat-q4_0.gguf",
"quant": "Q4_0",
"file_bytes": 6975877728,
"params_b": 12.0,
"japanese": true,
"notes": "Excellent quality/size balance, quantization-aware trained."
},
{
"name": "qwen3-14b",
"task": "chat",
"repo": "Qwen/Qwen3-14B-GGUF",
"filename": "Qwen3-14B-Q4_K_M.gguf",
"quant": "Q4_K_M",
"file_bytes": 9001752960,
"params_b": 14.0,
"japanese": true,
"notes": "High-quality dense model."
},
{
"name": "gemma-4-26b-a4b",
"task": "chat",
"repo": "google/gemma-4-26B-A4B-it-qat-q4_0-gguf",
"filename": "gemma-4-26B_q4_0-it.gguf",
"quant": "Q4_0",
"file_bytes": 14439361440,
"params_b": 26.0,
"japanese": true,
"notes": "MoE (4B active): near-flagship quality, fast even on CPU."
},
{
"name": "qwen3-32b",
"task": "chat",
"repo": "Qwen/Qwen3-32B-GGUF",
"filename": "Qwen3-32B-Q4_K_M.gguf",
"quant": "Q4_K_M",
"file_bytes": 19762149024,
"params_b": 32.0,
"japanese": true,
"notes": "Flagship dense quality; needs a big GPU or plenty of RAM."
},
{
"name": "qwen2.5-coder-7b",
"task": "code",
"repo": "Qwen/Qwen2.5-Coder-7B-Instruct-GGUF",
"filename": "qwen2.5-coder-7b-instruct-q4_k_m.gguf",
"quant": "Q4_K_M",
"file_bytes": 4683073536,
"params_b": 7.6,
"japanese": false,
"notes": "Solid small coding assistant."
},
{
"name": "qwen2.5-coder-14b",
"task": "code",
"repo": "Qwen/Qwen2.5-Coder-14B-Instruct-GGUF",
"filename": "qwen2.5-coder-14b-instruct-q4_k_m.gguf",
"quant": "Q4_K_M",
"file_bytes": 8988110272,
"params_b": 14.8,
"japanese": false,
"notes": "Mid-size coding assistant."
},
{
"name": "qwen3-coder-30b-a3b",
"task": "code",
"repo": "unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF",
"filename": "Qwen3-Coder-30B-A3B-Instruct-Q4_K_M.gguf",
"quant": "Q4_K_M",
"file_bytes": 18556689568,
"params_b": 30.5,
"japanese": false,
"notes": "MoE coder (3B active): top local coding quality, agent-capable."
},
{
"name": "ministral-3-3b-reasoning",
"task": "reasoning",
"repo": "mistralai/Ministral-3-3B-Reasoning-2512-GGUF",
"filename": "Ministral-3-3B-Reasoning-2512-Q4_K_M.gguf",
"quant": "Q4_K_M",
"file_bytes": 2147021472,
"params_b": 3.0,
"japanese": false,
"notes": "Small step-by-step reasoning model."
},
{
"name": "ministral-3-8b-reasoning",
"task": "reasoning",
"repo": "mistralai/Ministral-3-8B-Reasoning-2512-GGUF",
"filename": "Ministral-3-8B-Reasoning-2512-Q4_K_M.gguf",
"quant": "Q4_K_M",
"file_bytes": 5198910368,
"params_b": 8.0,
"japanese": false,
"notes": "Mid-size reasoning model."
},
{
"name": "ministral-3-14b-reasoning",
"task": "reasoning",
"repo": "mistralai/Ministral-3-14B-Reasoning-2512-GGUF",
"filename": "Ministral-3-14B-Reasoning-2512-Q4_K_M.gguf",
"quant": "Q4_K_M",
"file_bytes": 8239591488,
"params_b": 14.0,
"japanese": false,
"notes": "Strong math/logic reasoning."
},
{
"name": "gpt-oss-20b",
"task": "reasoning",
"repo": "ggml-org/gpt-oss-20b-GGUF",
"filename": "gpt-oss-20b-mxfp4.gguf",
"quant": "MXFP4",
"file_bytes": 12109566560,
"params_b": 20.9,
"japanese": false,
"notes": "OpenAI open-weight MoE (3.6B active); adjustable reasoning effort."
},
{
"name": "qwen3-embedding-0.6b",
"task": "embedding",
"repo": "Qwen/Qwen3-Embedding-0.6B-GGUF",
"filename": "Qwen3-Embedding-0.6B-Q8_0.gguf",
"quant": "Q8_0",
"file_bytes": 639150592,
"params_b": 0.6,
"japanese": true,
"notes": "Fast multilingual embeddings for RAG."
},
{
"name": "qwen3-embedding-4b",
"task": "embedding",
"repo": "Qwen/Qwen3-Embedding-4B-GGUF",
"filename": "Qwen3-Embedding-4B-Q8_0.gguf",
"quant": "Q8_0",
"file_bytes": 4279660224,
"params_b": 4.0,
"japanese": true,
"notes": "High-quality multilingual embeddings."
},
{
"name": "qwen3-embedding-8b",
"task": "embedding",
"repo": "Qwen/Qwen3-Embedding-8B-GGUF",
"filename": "Qwen3-Embedding-8B-Q8_0.gguf",
"quant": "Q8_0",
"file_bytes": 8047105824,
"params_b": 8.0,
"japanese": true,
"notes": "Top-tier retrieval quality."
},
{
"name": "whisper-base",
"task": "stt",
"format": "ggml",
"repo": "ggerganov/whisper.cpp",
"filename": "ggml-base.bin",
"quant": "F16",
"file_bytes": 147951465,
"params_b": 0.074,
"japanese": true,
"notes": "Fast speech-to-text for whisper.cpp; fine for quick notes."
},
{
"name": "whisper-small",
"task": "stt",
"format": "ggml",
"repo": "ggerganov/whisper.cpp",
"filename": "ggml-small.bin",
"quant": "F16",
"file_bytes": 487601967,
"params_b": 0.244,
"japanese": true,
"notes": "Good speed/accuracy balance for whisper.cpp."
},
{
"name": "whisper-large-v3-turbo",
"task": "stt",
"format": "ggml",
"repo": "ggerganov/whisper.cpp",
"filename": "ggml-large-v3-turbo.bin",
"quant": "F16",
"file_bytes": 1624555275,
"params_b": 0.809,
"japanese": true,
"notes": "Near-best accuracy at high speed for whisper.cpp."
},
{
"name": "outetts-0.3-500m",
"task": "tts",
"repo": "OuteAI/OuteTTS-0.3-500M-GGUF",
"filename": "OuteTTS-0.3-500M-Q8_0.gguf",
"quant": "Q8_0",
"file_bytes": 536666880,
"params_b": 0.5,
"japanese": true,
"notes": "Text-to-speech for the llama.cpp tts example (decoder included).",
"extra_files": [
{
"repo": "ggml-org/WavTokenizer",
"filename": "WavTokenizer-Large-75-F16.gguf",
"file_bytes": 130186688
}
]
},
{
"name": "outetts-0.3-1b",
"task": "tts",
"repo": "OuteAI/OuteTTS-0.3-1B-GGUF",
"filename": "OuteTTS-0.3-1B-Q8_0.gguf",
"quant": "Q8_0",
"file_bytes": 1267552832,
"params_b": 1.0,
"japanese": true,
"notes": "Higher-quality text-to-speech for the llama.cpp tts example (decoder included).",
"extra_files": [
{
"repo": "ggml-org/WavTokenizer",
"filename": "WavTokenizer-Large-75-F16.gguf",
"file_bytes": 130186688
}
]
},
{
"name": "gemma-4-e4b-vision",
"task": "vision",
"repo": "google/gemma-4-E4B-it-qat-q4_0-gguf",
"filename": "gemma-4-E4B_q4_0-it.gguf",
"quant": "Q4_0",
"file_bytes": 5154939136,
"params_b": 8.0,
"japanese": true,
"notes": "Image understanding + chat (vision projector included).",
"extra_files": [
{
"filename": "gemma-4-E4B-it-mmproj.gguf",
"file_bytes": 991551904
}
]
},
{
"name": "qwen3-vl-8b",
"task": "vision",
"repo": "Qwen/Qwen3-VL-8B-Instruct-GGUF",
"filename": "Qwen3VL-8B-Instruct-Q4_K_M.gguf",
"quant": "Q4_K_M",
"file_bytes": 5027784800,
"params_b": 8.0,
"japanese": true,
"notes": "Strong image/video understanding (vision projector included).",
"extra_files": [
{
"filename": "mmproj-Qwen3VL-8B-Instruct-Q8_0.gguf",
"file_bytes": 752289728
}
]
},
{
"name": "gemma-4-12b-vision",
"task": "vision",
"repo": "google/gemma-4-12B-it-qat-q4_0-gguf",
"filename": "gemma-4-12b-it-qat-q4_0.gguf",
"quant": "Q4_0",
"file_bytes": 6975877728,
"params_b": 12.0,
"japanese": true,
"notes": "High-quality image understanding + chat (vision projector included).",
"extra_files": [
{
"filename": "mmproj-gemma-4-12b-it-qat-q4_0.gguf",
"file_bytes": 175115264
}
]
}
]
}