@meta
  v: 1
  route: /model/mistral-small-3
  generated: 2026-06-10T08:49:35.883Z

@intent
  purpose:    Show what hardware can run Mistral Small 3.
  audience:   ai-engineer, self-hoster, model-evaluator
  capability: inspect_model_specs, find_compatible_gpus, compare_similar_models, open_huggingface

@state
  slug: mistral-small-3
  name: Mistral Small 3
  hf_repo: mistralai/Mistral-Small-24B-Instruct-2501
  params_b: 24
  active_params_b: ~
  family: Mistral
  type: dense
  context_k: 32
  fp16_gb: 48
  summary: January 2025 Mistral 24B. Llama 3.3 70B-tier quality at a third of the VRAM.
  smallest_fitting_gpu_q4: rtx-3090
  smallest_fitting_gpu_fp16: m2-max-64

@actions
  - id: open_huggingface
    method: GET
    href: https://huggingface.co/mistralai/Mistral-Small-24B-Instruct-2501
  - id: view_compatible_gpus
    method: GET
    href: /model/mistral-small-3
  - id: compare_similar
    method: GET
    href: /model/gemma-4-26b
  - id: view_calculator
    method: GET
    href: /#calculator

@context
  > January 2025 Mistral 24B. Llama 3.3 70B-tier quality at a third of the VRAM.

@nav
  self:      /model/mistral-small-3
  parents:   [/]
  peers:     [/model/gemma-4-26b, /model/qwen-3-6-27b, /model/gpt-oss-20b, /model/qwen-3-30b-a3b]
  drilldown: /the-math