{
 "id": "qwen2.5-14b-1m",
 "name": "Qwen2.5 14B-1M",
 "repo": "Qwen/Qwen2.5-14B-Instruct-1M",
 "lab": {
  "name": "Alibaba",
  "country": "China"
 },
 "released": "2025-01-23",
 "parameters": 14.770033664,
 "activeParameters": null,
 "maxContext": 1010000,
 "architecture": "48 layers · 48 full attention (8 KV × 128)",
 "page": "https://studiotvai.com/vram-requirements/qwen2-5-14b-1m",
 "calculator": "https://studiotvai.com/llm-vram-calculator?model=qwen2.5-14b-1m",
 "weightsGiB": {
  "bf16": 27.51,
  "fp8": 15.21,
  "q4_k_m": 8.46
 },
 "kvCacheGiBPerRequest": [
  {
   "context": 8192,
   "bf16": 1.5,
   "fp8": 0.75
  },
  {
   "context": 32768,
   "bf16": 6,
   "fp8": 3
  },
  {
   "context": 131072,
   "bf16": 24,
   "fp8": 12
  },
  {
   "context": 1010000,
   "bf16": 184.937,
   "fp8": 92.468
  }
 ],
 "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": 2,
     "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": "strix-halo",
    "name": "Ryzen AI Max+ 395 128 GB",
    "memoryGB": 96,
    "engine": "llama.cpp",
    "byFormat": {
     "bf16": 1,
     "fp8": 1,
     "q4_k_m": 1
    }
   },
   {
    "gpu": "dgx-spark",
    "name": "DGX Spark (GB10)",
    "memoryGB": 128,
    "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/Qwen2.5-14B-Instruct-1M-GGUF",
  "gib": {
   "IQ2_M": 4.99,
   "IQ2_S": 4.66,
   "IQ2_XS": 4.38,
   "IQ3_M": 6.44,
   "IQ3_XS": 5.94,
   "IQ4_NL": 7.96,
   "IQ4_XS": 7.56,
   "Q2_K": 5.37,
   "Q2_K_L": 6.08,
   "Q3_K_L": 7.38,
   "Q3_K_M": 6.84,
   "Q3_K_S": 6.2,
   "Q3_K_XL": 8.01,
   "Q4_0": 7.96,
   "Q4_1": 8.75,
   "Q4_K_L": 8.91,
   "Q4_K_M": 8.37,
   "Q4_K_S": 7.98,
   "Q5_K_L": 10.23,
   "Q5_K_M": 9.79,
   "Q5_K_S": 9.56,
   "Q6_K": 11.29,
   "Q6_K_L": 11.64,
   "Q8_0": 14.62,
   "F16": 27.52,
   "F32": 55.03
  }
 },
 "runsWith": [
  "vLLM",
  "SGLang",
  "llama.cpp (Ollama, LM Studio)",
  "MLX (Mac)"
 ],
 "verified": {
  "gguf": {
   "day": "2026-11-06",
   "repo": "bartowski/Qwen2.5-14B-Instruct-1M-GGUF",
   "kv": "match",
   "ours": 196608,
   "theirs": 196608
  }
 },
 "updated": "2026-11-06T21:02:26.190Z"
}