bennydaball/minimax-m2.5-reap-139b-a10b-gguf overview

{
  "metadata": {},
  "card_data": {
    "license": "other",
    "base_model": [
      "tomngdev/MiniMax-M2.5-REAP-139B-A10B-GGUF"
    ],
    "language": [
      "en"
    ],
    "tags": [
      "gguf",
      "minimax",
      "moe",
      "reap",
      "text-generation"
    ],
    "pipeline_tag": "text-generation",
    "frontmatter": {
      "license": "other",
      "base_model": [
        "tomngdev/MiniMax-M2.5-REAP-139B-A10B-GGUF"
      ],
      "language": [
        "en"
      ],
      "tags": [
        "gguf",
        "minimax",
        "moe",
        "reap",
        "text-generation"
      ],
      "pipeline_tag": "text-generation"
    },
    "hero_image_url": "",
    "summary": "This is the REAP model in practical pants: high quality GGUF quants for local inference without setting your workstation on fire. Built from:",
    "quick_links": [],
    "benchmark_table_html": "",
    "readme_markdown": "---\nlicense: other\nbase_model:\n- tomngdev/MiniMax-M2.5-REAP-139B-A10B-GGUF\nlanguage:\n- en\ntags:\n- gguf\n- minimax\n- moe\n- reap\n- text-generation\npipeline_tag: text-generation\n---\n\n# MiniMax-M2.5-REAP-139B-A10B-GGUF\n\nThis is the REAP model in practical pants: high quality GGUF quants for local inference without setting your workstation on fire.\n\nBuilt from:\n- Base: `MiniMaxAI/MiniMax-M2.5`\n- REAP source: `tomngdev/MiniMax-M2.5-REAP-139B-A10B-GGUF` (BF16 split)\n- Quantized locally with `llama.cpp` on Strix Halo + high RAM mode.\n\n## Available Quants\n\n| Quant | Status | Size (GiB) | Notes |\n|---|---|---:|---|\n| `Q8_0` | uploaded | 137.78 | Highest quality quant in this pack |\n| `Q5_K_M` | uploading | 92.33 | Better quality/size balance |\n| `Q4_K_M` | uploaded | 78.83 | Strong practical default |\n\n## File Layout\n\nAll quants are split GGUF sets (`00001-of-00007` etc.) for safer handling of very large models.\n\n## Quality Notes\n\n- These are generated from BF16 REAP GGUF, not requantized from lower precision.\n- Token embedding and output tensors are kept at `Q8_0` during quantization for quality retention.\n\n## Usage\n\nUse any first shard with `llama.cpp`; it auto-discovers sibling shards:\n\n```bash\nllama-cli -m MiniMax-M2.5-REAP-Q4_K_M-00001-of-00007.gguf -ngl 0 -c 8192\n```\n\n## Credits\n\n- `MiniMaxAI` for MiniMax-M2.5\n- `tomngdev` for the BF16 REAP GGUF release\n- `BennyDaBall` for this quant pack\n\n## Disclaimer\n\nYou are responsible for your own use, outputs, and compliance with applicable laws and platform policies.",
    "related_quantizations": []
  },
  "tags": [
    "gguf",
    "minimax",
    "moe",
    "reap",
    "text-generation",
    "en",
    "base_model:tomngdev/MiniMax-M2.5-REAP-139B-A10B-GGUF",
    "base_model:quantized:tomngdev/MiniMax-M2.5-REAP-139B-A10B-GGUF",
    "license:other",
    "endpoints_compatible",
    "region:us",
    "conversational"
  ],
  "likes": 4,
  "downloads": 176,
  "gated": false,
  "private": false,
  "last_modified": "2026-02-19T17:03:35.000Z",
  "created_at": "2026-02-19T03:58:16.000Z",
  "pipeline_tag": "text-generation",
  "library_name": ""
}

{
  "_id": "69968a58717d231f675a7007",
  "id": "BennyDaBall/MiniMax-M2.5-REAP-139B-A10B-GGUF",
  "modelId": "BennyDaBall/MiniMax-M2.5-REAP-139B-A10B-GGUF",
  "sha": "a9b07e70be737d99a5f7b8c002ea28373c15632f",
  "createdAt": "2026-02-19T03:58:16.000Z",
  "lastModified": "2026-02-19T17:03:35.000Z",
  "author": "BennyDaBall",
  "downloads": 176,
  "likes": 4,
  "gated": false,
  "private": false,
  "pipeline_tag": "text-generation",
  "library_name": "",
  "siblings_count": 23
}