{
 "id": "mistral-small-3.2",
 "name": "Mistral Small 3.2 24B",
 "repo": "mistralai/Mistral-Small-3.2-24B-Instruct-2506",
 "lab": {
  "name": "Mistral AI",
  "country": "France"
 },
 "released": "2025-06-19",
 "parameters": 24.01,
 "activeParameters": null,
 "maxContext": 131072,
 "architecture": "40 layers · GQA, 8 KV heads × 128 · checkpoint stored in FP32, served in BF16",
 "page": "https://studiotvai.com/vram-requirements/mistral-small-3-2",
 "calculator": "https://studiotvai.com/llm-vram-calculator?model=mistral-small-3.2",
 "weightsGiB": {
  "bf16": 44.72,
  "fp8": 23.61,
  "q4_k_m": 13.39
 },
 "kvCacheGiBPerRequest": [
  {
   "context": 8192,
   "bf16": 1.25,
   "fp8": 0.625
  },
  {
   "context": 32768,
   "bf16": 5,
   "fp8": 2.5
  },
  {
   "context": 131072,
   "bf16": 20,
   "fp8": 10
  }
 ],
 "recurrentStateGiBPerRequest": 0,
 "gpusNeeded": {
  "scenario": "one request of 8192 tokens, BF16 KV cache; vLLM on data-center GPUs, llama.cpp elsewhere",
  "gpus": [
   {
    "gpu": "rtx-4090",
    "name": "GeForce RTX 4090",
    "memoryGB": 24,
    "engine": "llama.cpp",
    "byFormat": {
     "bf16": 3,
     "fp8": 2,
     "q4_k_m": 1
    }
   },
   {
    "gpu": "rtx-5090",
    "name": "GeForce RTX 5090",
    "memoryGB": 32,
    "engine": "llama.cpp",
    "byFormat": {
     "bf16": 2,
     "fp8": 1,
     "q4_k_m": 1
    }
   },
   {
    "gpu": "rtx-pro-6000",
    "name": "RTX PRO 6000 Blackwell",
    "memoryGB": 96,
    "engine": "llama.cpp",
    "byFormat": {
     "bf16": 1,
     "fp8": 1,
     "q4_k_m": 1
    }
   },
   {
    "gpu": "mac-m5-max-128",
    "name": "Mac M5 Max 128 GB",
    "memoryGB": 96,
    "engine": "llama.cpp",
    "byFormat": {
     "bf16": 1,
     "fp8": 1,
     "q4_k_m": 1
    }
   },
   {
    "gpu": "h100",
    "name": "NVIDIA H100",
    "memoryGB": 80,
    "engine": "vLLM",
    "byFormat": {
     "bf16": 1,
     "fp8": 1,
     "q4_k_m": 1
    }
   },
   {
    "gpu": "h200",
    "name": "NVIDIA H200",
    "memoryGB": 141,
    "engine": "vLLM",
    "byFormat": {
     "bf16": 1,
     "fp8": 1,
     "q4_k_m": 1
    }
   },
   {
    "gpu": "b200",
    "name": "NVIDIA B200 (HGX)",
    "memoryGB": 180,
    "engine": "vLLM",
    "byFormat": {
     "bf16": 1,
     "fp8": 1,
     "q4_k_m": 1
    }
   }
  ]
 },
 "ggufFiles": {
  "repo": "unsloth/Mistral-Small-3.2-24B-Instruct-2506-GGUF",
  "gib": {
   "BF16": 43.92,
   "IQ4_NL": 12.54,
   "IQ4_XS": 11.9,
   "Q2_K": 8.28,
   "Q2_K_L": 8.43,
   "Q3_K_M": 10.69,
   "Q3_K_S": 9.69,
   "Q4_0": 12.57,
   "Q4_1": 13.85,
   "Q4_K_M": 13.35,
   "Q4_K_S": 12.62,
   "Q5_K_M": 15.61,
   "Q5_K_S": 15.18,
   "Q6_K": 18.02,
   "Q8_0": 23.33,
   "UD-IQ1_M": 5.6,
   "UD-IQ1_S": 5.18,
   "UD-IQ2_M": 7.68,
   "UD-IQ2_XXS": 6.29,
   "UD-IQ3_XXS": 8.76,
   "UD-Q2_K_XL": 8.65,
   "UD-Q3_K_XL": 11.04,
   "UD-Q4_K_XL": 13.55,
   "UD-Q5_K_XL": 15.64,
   "UD-Q6_K_XL": 19.36,
   "UD-Q8_K_XL": 27
  }
 },
 "runsWith": [
  "vLLM",
  "SGLang",
  "llama.cpp (Ollama, LM Studio)",
  "MLX (Mac)"
 ],
 "verified": {
  "gguf": {
   "day": "2026-10-10",
   "repo": "unsloth/Mistral-Small-3.2-24B-Instruct-2506-GGUF",
   "kv": "match",
   "ours": 163840,
   "theirs": 163840
  }
 },
 "updated": "2026-10-10T15:05:26.095Z"
}