{
 "id": "kimi-k3.1-pro",
 "name": "Kimi K3.1 Pro",
 "repo": "moonshotai/Kimi-K3.1-Pro",
 "lab": {
  "name": "Moonshot AI",
  "country": "China"
 },
 "released": "2026-11-04",
 "parameters": 2780.110363392,
 "activeParameters": 106,
 "maxContext": 1048576,
 "architecture": "93 layers · MLA (512 + 64 latent) on 24 · 69 Kimi Delta Attention",
 "page": "https://studiotvai.com/vram-requirements/kimi-k3-1-pro",
 "calculator": "https://studiotvai.com/llm-vram-calculator?model=kimi-k3.1-pro",
 "weightsGiB": {
  "native": 1454.07,
  "bf16": 5178.36,
  "fp8": 2591.37,
  "q4_k_m": 1563.45
 },
 "kvCacheGiBPerRequest": [
  {
   "context": 8192,
   "bf16": 0.211,
   "fp8": 0.105
  },
  {
   "context": 32768,
   "bf16": 0.844,
   "fp8": 0.422
  },
  {
   "context": 131072,
   "bf16": 3.375,
   "fp8": 1.688
  },
  {
   "context": 1048576,
   "bf16": 27,
   "fp8": 13.5
  }
 ],
 "recurrentStateGiBPerRequest": 0.4185,
 "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": 96,
     "bf16": null,
     "fp8": null,
     "q4_k_m": 96
    }
   },
   {
    "gpu": "rtx-5090",
    "name": "GeForce RTX 5090",
    "memoryGB": 32,
    "engine": "llama.cpp",
    "byFormat": {
     "native": 64,
     "bf16": null,
     "fp8": 96,
     "q4_k_m": 64
    }
   },
   {
    "gpu": "rtx-pro-6000",
    "name": "RTX PRO 6000 Blackwell",
    "memoryGB": 96,
    "engine": "llama.cpp",
    "byFormat": {
     "native": 16,
     "bf16": 64,
     "fp8": 32,
     "q4_k_m": 24
    }
   },
   {
    "gpu": "mac-m5-max-128",
    "name": "Mac M5 Max 128 GB",
    "memoryGB": 96,
    "engine": "llama.cpp",
    "byFormat": {
     "native": 16,
     "bf16": 64,
     "fp8": 32,
     "q4_k_m": 24
    }
   },
   {
    "gpu": "strix-halo",
    "name": "Ryzen AI Max+ 395 128 GB",
    "memoryGB": 96,
    "engine": "llama.cpp",
    "byFormat": {
     "native": 16,
     "bf16": 64,
     "fp8": 32,
     "q4_k_m": 24
    }
   },
   {
    "gpu": "dgx-spark",
    "name": "DGX Spark (GB10)",
    "memoryGB": 128,
    "engine": "llama.cpp",
    "byFormat": {
     "native": 16,
     "bf16": 48,
     "fp8": 24,
     "q4_k_m": 16
    }
   },
   {
    "gpu": "h100",
    "name": "NVIDIA H100",
    "memoryGB": 80,
    "engine": "vLLM",
    "byFormat": {
     "native": 24,
     "bf16": 96,
     "fp8": 48,
     "q4_k_m": 24
    }
   },
   {
    "gpu": "h200",
    "name": "NVIDIA H200",
    "memoryGB": 141,
    "engine": "vLLM",
    "byFormat": {
     "native": 16,
     "bf16": 48,
     "fp8": 24,
     "q4_k_m": 16
    }
   },
   {
    "gpu": "b200",
    "name": "NVIDIA B200 (HGX)",
    "memoryGB": 180,
    "engine": "vLLM",
    "byFormat": {
     "native": 16,
     "bf16": 48,
     "fp8": 24,
     "q4_k_m": 16
    }
   }
  ]
 },
 "ggufFiles": null,
 "runsWith": [
  "SGLang"
 ],
 "verified": null,
 "updated": "2026-11-05T00:08:05.925Z"
}