@meta
  v: 1
  route: /model/mixtral-8x22b
  generated: 2026-08-01T19:27:22.576Z

@intent
  purpose:    Show what hardware can run Mixtral 8x22B.
  audience:   ai-engineer, self-hoster, model-evaluator
  capability: inspect_model_specs, find_compatible_gpus, compare_similar_models, open_huggingface

@state
  slug: mixtral-8x22b
  name: Mixtral 8x22B
  hf_repo: mistralai/Mixtral-8x22B-Instruct-v0.1
  params_b: 141
  active_params_b: 39
  family: Mistral
  type: moe
  context_k: 64
  fp16_gb: 282
  summary: "Big Mixtral. 141B total, 39B active. Multi-GPU or M-series Ultra territory."
  smallest_fitting_gpu_q4: m4-max-128
  smallest_fitting_gpu_fp16: m3-ultra-512

@actions
  - id: open_huggingface
    method: GET
    href: https://huggingface.co/mistralai/Mixtral-8x22B-Instruct-v0.1
  - id: view_compatible_gpus
    method: GET
    href: /model/mixtral-8x22b
  - id: compare_similar
    method: GET
    href: /model/gpt-oss-120b
  - id: view_calculator
    method: GET
    href: /#calculator

@context
  > Big Mixtral. 141B total, 39B active. Multi-GPU or M-series Ultra territory.

@nav
  self:      /model/mixtral-8x22b
  parents:   [/]
  peers:     [/model/gpt-oss-120b, /model/command-r-plus, /model/qwen-2-5-72b, /model/llama-3-3-70b]
  drilldown: /the-math