{
 "id": "mistral-large-3",
 "name": "Mistral Large 3",
 "repo": "mistralai/Mistral-Large-3-675B-Instruct-2512",
 "lab": {
  "name": "Mistral AI",
  "country": "France"
 },
 "released": "2025-12-02",
 "parameters": 675.99,
 "activeParameters": 41,
 "maxContext": 262144,
 "architecture": "MoE 128 experts + 1 shared · MLA (512 + 64 latent) on 61 layers · vision encoder",
 "page": "https://studiotvai.com/vram-requirements/mistral-large-3",
 "calculator": "https://studiotvai.com/llm-vram-calculator?model=mistral-large-3",
 "weightsGiB": {
  "native": 634.7,
  "bf16": 1259.13,
  "fp8": 631.31,
  "q4_k_m": 378.84
 },
 "kvCacheGiBPerRequest": [
  {
   "context": 8192,
   "bf16": 0.536,
   "fp8": 0.268
  },
  {
   "context": 32768,
   "bf16": 2.145,
   "fp8": 1.072
  },
  {
   "context": 131072,
   "bf16": 8.578,
   "fp8": 4.289
  },
  {
   "context": 262144,
   "bf16": 17.156,
   "fp8": 8.578
  }
 ],
 "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": {
     "native": 32,
     "bf16": 64,
     "fp8": 32,
     "q4_k_m": 24
    }
   },
   {
    "gpu": "rtx-5090",
    "name": "GeForce RTX 5090",
    "memoryGB": 32,
    "engine": "llama.cpp",
    "byFormat": {
     "native": 24,
     "bf16": 48,
     "fp8": 24,
     "q4_k_m": 16
    }
   },
   {
    "gpu": "rtx-pro-6000",
    "name": "RTX PRO 6000 Blackwell",
    "memoryGB": 96,
    "engine": "llama.cpp",
    "byFormat": {
     "native": 7,
     "bf16": 16,
     "fp8": 7,
     "q4_k_m": 5
    }
   },
   {
    "gpu": "mac-m5-max-128",
    "name": "Mac M5 Max 128 GB",
    "memoryGB": 96,
    "engine": "llama.cpp",
    "byFormat": {
     "native": 7,
     "bf16": 16,
     "fp8": 7,
     "q4_k_m": 5
    }
   },
   {
    "gpu": "strix-halo",
    "name": "Ryzen AI Max+ 395 128 GB",
    "memoryGB": 96,
    "engine": "llama.cpp",
    "byFormat": {
     "native": 7,
     "bf16": 16,
     "fp8": 7,
     "q4_k_m": 5
    }
   },
   {
    "gpu": "dgx-spark",
    "name": "DGX Spark (GB10)",
    "memoryGB": 128,
    "engine": "llama.cpp",
    "byFormat": {
     "native": 6,
     "bf16": 16,
     "fp8": 6,
     "q4_k_m": 4
    }
   },
   {
    "gpu": "h100",
    "name": "NVIDIA H100",
    "memoryGB": 80,
    "engine": "vLLM",
    "byFormat": {
     "native": 16,
     "bf16": 24,
     "fp8": 16,
     "q4_k_m": 6
    }
   },
   {
    "gpu": "h200",
    "name": "NVIDIA H200",
    "memoryGB": 141,
    "engine": "vLLM",
    "byFormat": {
     "native": 6,
     "bf16": 16,
     "fp8": 6,
     "q4_k_m": 4
    }
   },
   {
    "gpu": "b200",
    "name": "NVIDIA B200 (HGX)",
    "memoryGB": 180,
    "engine": "vLLM",
    "byFormat": {
     "native": 5,
     "bf16": 16,
     "fp8": 5,
     "q4_k_m": 3
    }
   }
  ]
 },
 "ggufFiles": {
  "repo": "unsloth/Mistral-Large-3-675B-Instruct-2512-GGUF",
  "gib": {
   "BF16": 1254.45,
   "IQ4_NL": 354.59,
   "IQ4_XS": 335.23,
   "UD-TQ1_0": 157.03,
   "Q2_K": 230.14,
   "Q2_K_L": 230.34,
   "Q3_K_M": 299.74,
   "Q3_K_S": 271.81,
   "Q4_0": 355.49,
   "Q4_1": 393.36,
   "Q4_K_M": 379.04,
   "Q4_K_S": 356.38,
   "Q5_K_M": 445.14,
   "Q5_K_S": 432.55,
   "Q6_K": 515.32,
   "Q8_0": 666.53,
   "UD-IQ1_M": 192.28,
   "UD-IQ1_S": 177.7,
   "UD-IQ2_M": 219.21,
   "UD-IQ2_XXS": 209.37,
   "UD-IQ3_XXS": 260.04,
   "UD-Q2_K_XL": 238.76,
   "UD-Q3_K_XL": 280.13,
   "UD-Q4_K_XL": 361.24,
   "UD-Q5_K_XL": 446.85,
   "UD-Q6_K_XL": 536.88,
   "UD-Q8_K_XL": 720.38
  }
 },
 "runsWith": [
  "vLLM"
 ],
 "verified": null,
 "updated": "2026-10-30T00:07:05.915Z"
}