@meta
  v: 1
  route: /model/qwen-2-5-32b
  generated: 2026-06-10T09:07:23.050Z

@intent
  purpose:    Show what hardware can run Qwen 2.5 32B.
  audience:   ai-engineer, self-hoster, model-evaluator
  capability: inspect_model_specs, find_compatible_gpus, compare_similar_models, open_huggingface

@state
  slug: qwen-2-5-32b
  name: Qwen 2.5 32B
  hf_repo: Qwen/Qwen2.5-32B-Instruct
  params_b: 32.5
  active_params_b: ~
  family: Alibaba
  type: dense
  context_k: 128
  fp16_gb: 65
  summary: "Alibaba's mid-range workhorse. Q4_K_M lands at ~18 GB, the canonical \"fits on a 24GB\" 30B-class."
  smallest_fitting_gpu_q4: rtx-3090
  smallest_fitting_gpu_fp16: h100

@actions
  - id: open_huggingface
    method: GET
    href: https://huggingface.co/Qwen/Qwen2.5-32B-Instruct
  - id: view_compatible_gpus
    method: GET
    href: /model/qwen-2-5-32b
  - id: compare_similar
    method: GET
    href: /model/qwen-2-5-coder-32b
  - id: view_calculator
    method: GET
    href: /#calculator

@context
  > Alibaba's mid-range workhorse. Q4_K_M lands at ~18 GB, the canonical "fits on a 24GB" 30B-class.

@nav
  self:      /model/qwen-2-5-32b
  parents:   [/]
  peers:     [/model/qwen-2-5-coder-32b, /model/yi-34b, /model/qwen-3-30b-a3b, /model/qwen-3-6-35b-a3b]
  drilldown: /the-math