{
  "hf_id": "XiaomiMiMo/MiMo-V2.5-Pro",
  "meta": {
    "title": "MiMo-V2.5-Pro",
    "slug": "mimo-v2-5-pro",
    "provider": "MiMo (Xiaomi)",
    "description": "Xiaomi's flagship MoE reasoning model (1.02T total / 42B active) with hybrid attention, native FP8 weights, and Multi-Token Prediction",
    "date_added": "2026-04-27",
    "date_updated": "2026-04-27",
    "tasks": [
      "text"
    ],
    "related_recipes": [
      "XiaomiMiMo/MiMo-V2.5"
    ],
    "hardware": {
      "h200": "verified"
    }
  },
  "recommended_command": {
    "hardware": "h200",
    "strategy": "single_node_tp",
    "variant": "default",
    "node_count": 1,
    "deploy_type": "single_node",
    "env": {},
    "docker_image": "vllm/vllm-openai:latest",
    "command": "vllm serve XiaomiMiMo/MiMo-V2.5-Pro \\\n  --trust-remote-code \\\n  --max-model-len auto \\\n  --generation-config vllm \\\n  --gpu-memory-utilization 0.95 \\\n  --max-model-len auto \\\n  --tensor-parallel-size 8 \\\n  --tool-call-parser mimo \\\n  --enable-auto-tool-choice \\\n  --reasoning-parser mimo",
    "argv": [
      "vllm",
      "serve",
      "XiaomiMiMo/MiMo-V2.5-Pro",
      "--trust-remote-code",
      "--max-model-len auto",
      "--generation-config",
      "vllm",
      "--gpu-memory-utilization",
      "0.95",
      "--max-model-len",
      "auto",
      "--tensor-parallel-size",
      "8",
      "--tool-call-parser",
      "mimo",
      "--enable-auto-tool-choice",
      "--reasoning-parser",
      "mimo"
    ],
    "docker_command": "docker run --gpus all \\\n  --privileged --ipc=host -p 8000:8000 \\\n  -v ~/.cache/huggingface:/root/.cache/huggingface \\\n  vllm/vllm-openai:latest XiaomiMiMo/MiMo-V2.5-Pro \\\n  --trust-remote-code \\\n  --max-model-len auto \\\n  --generation-config vllm \\\n  --gpu-memory-utilization 0.95 \\\n  --max-model-len auto \\\n  --tensor-parallel-size 8 \\\n  --tool-call-parser mimo \\\n  --enable-auto-tool-choice \\\n  --reasoning-parser mimo",
    "docker_argv": [
      "docker",
      "run",
      "--gpus",
      "all",
      "--privileged",
      "--ipc=host",
      "-p",
      "8000:8000",
      "-v",
      "~/.cache/huggingface:/root/.cache/huggingface",
      "vllm/vllm-openai:latest",
      "XiaomiMiMo/MiMo-V2.5-Pro",
      "--trust-remote-code",
      "--max-model-len auto",
      "--generation-config",
      "vllm",
      "--gpu-memory-utilization",
      "0.95",
      "--max-model-len",
      "auto",
      "--tensor-parallel-size",
      "8",
      "--tool-call-parser",
      "mimo",
      "--enable-auto-tool-choice",
      "--reasoning-parser",
      "mimo"
    ],
    "strategy_spec": {
      "name": "single_node_tp",
      "deploy_type": "single_node",
      "display_name": "Tensor Parallel",
      "short_label": "TP",
      "orientation": "latency",
      "description": "Single-node tensor parallel. Splits the model across all local GPUs. TP size is set to the GPU count at deploy time. The simplest multi-GPU strategy — works for all model architectures.\n",
      "hardware_match": {
        "min_gpus": 1,
        "max_gpus": 8,
        "multi_node": false
      },
      "vllm_args": [],
      "parallel_flag": "--tensor-parallel-size"
    },
    "hardware_profile": {
      "brand": "NVIDIA",
      "generation": "hopper",
      "compute_arch": "sm90",
      "display_name": "H200",
      "description": "NVIDIA H200 SXM 141 GB HBM3e · 8-GPU HGX node",
      "gpu_count": 8,
      "vram_gb": 1128,
      "multi_node": false
    },
    "alternatives": {
      "single_node_tep": "/XiaomiMiMo/MiMo-V2.5-Pro/strategies/single_node_tep.json",
      "pd_cluster": "/XiaomiMiMo/MiMo-V2.5-Pro/strategies/pd_cluster.json",
      "kv_store_centralized_mooncake": "/XiaomiMiMo/MiMo-V2.5-Pro/strategies/kv_store_centralized_mooncake.json",
      "kv_store_distributed_mooncake": "/XiaomiMiMo/MiMo-V2.5-Pro/strategies/kv_store_distributed_mooncake.json"
    },
    "by_hardware": {
      "h200": "/XiaomiMiMo/MiMo-V2.5-Pro/hw/h200.json",
      "h100": "/XiaomiMiMo/MiMo-V2.5-Pro/hw/h100.json",
      "b200": "/XiaomiMiMo/MiMo-V2.5-Pro/hw/b200.json",
      "gb200": "/XiaomiMiMo/MiMo-V2.5-Pro/hw/gb200.json",
      "b300": "/XiaomiMiMo/MiMo-V2.5-Pro/hw/b300.json",
      "gb300": "/XiaomiMiMo/MiMo-V2.5-Pro/hw/gb300.json",
      "mi300x": "/XiaomiMiMo/MiMo-V2.5-Pro/hw/mi300x.json",
      "mi325x": "/XiaomiMiMo/MiMo-V2.5-Pro/hw/mi325x.json",
      "mi350x": "/XiaomiMiMo/MiMo-V2.5-Pro/hw/mi350x.json",
      "mi355x": "/XiaomiMiMo/MiMo-V2.5-Pro/hw/mi355x.json"
    }
  },
  "model": {
    "model_id": "XiaomiMiMo/MiMo-V2.5-Pro",
    "min_vllm_version": "0.21.0",
    "architecture": "moe",
    "parameter_count": "1T",
    "active_parameters": "42B",
    "context_length": 1048576,
    "base_args": [
      "--trust-remote-code",
      "--max-model-len auto",
      "--generation-config",
      "vllm"
    ],
    "base_env": {},
    "install": {
      "pip": {
        "command": "uv venv\nsource .venv/bin/activate\nuv pip install -U vllm --torch-backend auto"
      },
      "docker": {
        "command": "docker pull vllm/vllm-openai:latest"
      }
    }
  },
  "features": {
    "tool_calling": {
      "description": "MiMo tool-call parser",
      "args": [
        "--tool-call-parser",
        "mimo",
        "--enable-auto-tool-choice"
      ]
    },
    "reasoning": {
      "description": "MiMo reasoning parser",
      "args": [
        "--reasoning-parser",
        "mimo"
      ]
    },
    "spec_decoding": {
      "description": "Multi-Token Prediction",
      "args": [
        "--speculative-config",
        "{\"method\":\"mtp\",\"num_speculative_tokens\":1}"
      ]
    }
  },
  "opt_in_features": [
    "spec_decoding"
  ],
  "variants": {
    "default": {
      "precision": "fp8",
      "vram_minimum_gb": 1224,
      "description": "Native FP8 weights (block-wise e4m3 128x128); 8x H200 with TP8",
      "extra_args": [
        "--tensor-parallel-size",
        "8",
        "--gpu-memory-utilization",
        "0.95",
        "--max-model-len",
        "auto"
      ]
    }
  },
  "hardware_overrides": {},
  "guide": "## Overview\n\nMiMo-V2.5-Pro is Xiaomi's flagship MoE reasoning model with 1.02T total parameters\nand 42B active per token. It uses 384 routed experts (top-8) with hybrid\nattention (full-attention + sliding-window 128 at 6:1 ratio) over 70 layers\n(1 dense + 69 MoE) and ships with native FP8 (block-wise e4m3) weights.\nA 3-layer Multi-Token Prediction (MTP) head enables speculative decoding for\n~3x output speed.\n\n## Prerequisites\n\n- Hardware: 8x H200 (TP8)\n\n### Pull the vLLM docker image\n\nStable vLLM does not yet support MiMo V2.5. Use the pre-built image:\n\n```bash\ndocker pull vllm/vllm-openai:mimov25-cu129\n```\n\n## Launch commands\n\nSingle-node TP8 (H200):\n\n```bash\nvllm serve XiaomiMiMo/MiMo-V2.5-Pro \\\n  --tensor-parallel-size 8 \\\n  --trust-remote-code \\\n  --gpu-memory-utilization 0.95 \\\n  --max-model-len auto \\\n  --generation-config vllm\n```\n\nWith tool calling + reasoning:\n\n```bash\nvllm serve XiaomiMiMo/MiMo-V2.5-Pro \\\n  --tensor-parallel-size 8 \\\n  --trust-remote-code \\\n  --gpu-memory-utilization 0.95 \\\n  --max-model-len auto \\\n  --reasoning-parser mimo \\\n  --tool-call-parser mimo \\\n  --enable-auto-tool-choice \\\n  --generation-config vllm\n```\n\n## Client Usage\n\n```bash\ncurl -X POST http://localhost:8000/v1/chat/completions \\\n  -H \"Content-Type: application/json\" \\\n  -d '{\n    \"model\": \"XiaomiMiMo/MiMo-V2.5-Pro\",\n    \"messages\": [{\"role\": \"user\", \"content\": \"Hello MiMo!\"}],\n    \"chat_template_kwargs\": {\"enable_thinking\": true}\n  }'\n```\n\nSet `\"enable_thinking\": false` (or omit the kwargs) to disable thinking mode.\n\n\n## References\n\n- [MiMo-V2.5-Pro on Hugging Face](https://huggingface.co/XiaomiMiMo/MiMo-V2.5-Pro)\n",
  "source": {
    "source": "vllm",
    "repo": "vllm-project/vllm",
    "ref": "main",
    "sha": "564ac0de1b228568cafcbfbe8ccaccc82d6023eb",
    "path": "recipes/models/XiaomiMiMo/MiMo-V2.5-Pro.yaml",
    "edit_url": "https://github.com/vllm-project/vllm/edit/main/recipes/models/XiaomiMiMo/MiMo-V2.5-Pro.yaml",
    "overlays": []
  }
}