{
 "id": "deepseek-r1-distill-llama-8b",
 "name": "DeepSeek-R1-Distill-Llama 8B",
 "repo": "deepseek-ai/DeepSeek-R1-Distill-Llama-8B",
 "lab": {
  "name": "DeepSeek",
  "country": "China"
 },
 "released": "2025-01-20",
 "parameters": 8.030261248,
 "activeParameters": null,
 "maxContext": 131072,
 "architecture": "32 layers · 32 full attention (8 KV × 128)",
 "page": "https://studiotvai.com/vram-requirements/deepseek-r1-distill-llama-8b",
 "calculator": "https://studiotvai.com/llm-vram-calculator?model=deepseek-r1-distill-llama-8b",
 "weightsGiB": {
  "bf16": 14.96,
  "fp8": 8.46,
  "q4_k_m": 4.62
 },
 "kvCacheGiBPerRequest": [
  {
   "context": 8192,
   "bf16": 1,
   "fp8": 0.5
  },
  {
   "context": 32768,
   "bf16": 4,
   "fp8": 2
  },
  {
   "context": 131072,
   "bf16": 16,
   "fp8": 8
  }
 ],
 "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": "unsloth/DeepSeek-R1-Distill-Llama-8B-GGUF",
  "gib": {
   "BF16": 14.97,
   "F16": 14.97,
   "IQ4_NL": 4.36,
   "IQ4_XS": 4.16,
   "Q2_K": 2.96,
   "Q2_K_L": 3.08,
   "Q3_K_M": 3.74,
   "Q3_K_S": 3.41,
   "Q4_0": 4.35,
   "Q4_1": 4.78,
   "Q4_K_M": 4.58,
   "Q4_K_S": 4.37,
   "Q5_K_M": 5.34,
   "Q5_K_S": 5.21,
   "Q6_K": 6.14,
   "Q8_0": 7.95,
   "UD-IQ1_M": 2.13,
   "UD-IQ1_S": 2.02,
   "UD-IQ2_M": 2.8,
   "UD-IQ2_XXS": 2.33,
   "UD-IQ3_XXS": 3.09,
   "UD-Q2_K_XL": 3.16,
   "UD-Q3_K_XL": 3.91,
   "UD-Q4_K_XL": 4.65,
   "UD-Q5_K_XL": 5.34,
   "UD-Q6_K_XL": 6.83,
   "UD-Q8_K_XL": 9.85
  }
 },
 "runsWith": [
  "vLLM",
  "SGLang",
  "llama.cpp (Ollama, LM Studio)",
  "MLX (Mac)"
 ],
 "verified": {
  "gguf": {
   "day": "2026-10-11",
   "repo": "unsloth/DeepSeek-R1-Distill-Llama-8B-GGUF",
   "kv": "match",
   "ours": 131072,
   "theirs": 131072
  }
 },
 "updated": "2026-10-11T19:08:05.666Z"
}