{
 "id": "ministral-8b",
 "name": "Ministral 8B",
 "repo": "mistralai/Ministral-8B-Instruct-2410",
 "lab": {
  "name": "Mistral AI",
  "country": "France"
 },
 "released": "2024-10-15",
 "parameters": 8.019808256,
 "activeParameters": null,
 "maxContext": 32768,
 "architecture": "36 layers · 9 full attention (8 KV × 128) · 27 sliding-window layers (32768 tokens, 8 KV × 128)",
 "page": "https://studiotvai.com/vram-requirements/ministral-8b",
 "calculator": "https://studiotvai.com/llm-vram-calculator?model=ministral-8b",
 "weightsGiB": {
  "bf16": 14.94,
  "fp8": 8.47,
  "q4_k_m": 4.62
 },
 "kvCacheGiBPerRequest": [
  {
   "context": 8192,
   "bf16": 1.125,
   "fp8": 0.563
  },
  {
   "context": 32768,
   "bf16": 4.5,
   "fp8": 2.25
  }
 ],
 "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": 1,
     "fp8": 1,
     "q4_k_m": 1
    }
   },
   {
    "gpu": "rtx-5090",
    "name": "GeForce RTX 5090",
    "memoryGB": 32,
    "engine": "llama.cpp",
    "byFormat": {
     "bf16": 1,
     "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": "bartowski/Ministral-8B-Instruct-2410-GGUF",
  "gib": {
   "IQ2_M": 2.76,
   "IQ3_M": 3.53,
   "IQ3_XS": 3.28,
   "IQ4_XS": 4.14,
   "Q2_K": 2.97,
   "Q2_K_L": 3.45,
   "Q3_K_L": 4.03,
   "Q3_K_M": 3.74,
   "Q3_K_S": 3.41,
   "Q3_K_XL": 4.47,
   "Q4_0": 4.35,
   "Q4_0_4_4": 4.34,
   "Q4_0_4_8": 4.34,
   "Q4_0_8_8": 4.34,
   "Q4_K_L": 4.95,
   "Q4_K_M": 4.57,
   "Q4_K_S": 4.36,
   "Q5_K_L": 5.64,
   "Q5_K_M": 5.33,
   "Q5_K_S": 5.21,
   "Q6_K": 6.14,
   "Q6_K_L": 6.38,
   "Q8_0": 7.94,
   "F16": 14.95
  }
 },
 "runsWith": [
  "vLLM",
  "SGLang",
  "llama.cpp (Ollama, LM Studio)",
  "MLX (Mac)"
 ],
 "verified": null,
 "updated": "2026-10-11T19:08:05.666Z"
}