@meta
  v: 1
  route: /model/qwen-3-6-35b-a3b
  generated: 2026-06-10T08:57:13.505Z

@intent
  purpose:    Show what hardware can run Qwen 3.6 35B A3B.
  audience:   ai-engineer, self-hoster, model-evaluator
  capability: inspect_model_specs, find_compatible_gpus, compare_similar_models, open_huggingface

@state
  slug: qwen-3-6-35b-a3b
  name: Qwen 3.6 35B A3B
  hf_repo: Qwen/Qwen3.6-35B-A3B
  params_b: 35
  active_params_b: 3
  family: Alibaba
  type: moe
  context_k: 256
  fp16_gb: 70
  summary: "Q2 2026 Qwen3.6 MoE flagship. 35B total, 3B active. Vision + multilingual + tool calling."
  smallest_fitting_gpu_q4: rtx-5090
  smallest_fitting_gpu_fp16: m3-max-96

@actions
  - id: open_huggingface
    method: GET
    href: https://huggingface.co/Qwen/Qwen3.6-35B-A3B
  - id: view_compatible_gpus
    method: GET
    href: /model/qwen-3-6-35b-a3b
  - id: compare_similar
    method: GET
    href: /model/yi-34b
  - id: view_calculator
    method: GET
    href: /#calculator

@context
  > Q2 2026 Qwen3.6 MoE flagship. 35B total, 3B active. Vision + multilingual + tool calling.

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