@meta
  v: 1
  route: /model/gpt-oss-120b
  generated: 2026-06-10T09:07:23.053Z

@intent
  purpose:    Show what hardware can run gpt-oss 120B.
  audience:   ai-engineer, self-hoster, model-evaluator
  capability: inspect_model_specs, find_compatible_gpus, compare_similar_models, open_huggingface

@state
  slug: gpt-oss-120b
  name: gpt-oss 120B
  hf_repo: openai/gpt-oss-120b
  params_b: 117
  active_params_b: 5.1
  family: OpenAI
  type: moe
  context_k: 128
  fp16_gb: 234
  summary: OpenAI's open-weight flagship. Near o4-mini on core reasoning; MXFP4 deployment fits a single 80GB H100.
  smallest_fitting_gpu_q4: m3-max-96
  smallest_fitting_gpu_fp16: m3-ultra-512

@actions
  - id: open_huggingface
    method: GET
    href: https://huggingface.co/openai/gpt-oss-120b
  - id: view_compatible_gpus
    method: GET
    href: /model/gpt-oss-120b
  - id: compare_similar
    method: GET
    href: /model/command-r-plus
  - id: view_calculator
    method: GET
    href: /#calculator

@context
  > OpenAI's open-weight flagship. Near o4-mini on core reasoning; MXFP4 deployment fits a single 80GB H100.

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