{
 "id": "nemotron-3.5-bolt-74b-a9b",
 "name": "Nemotron 3.5 Bolt 74B-A9B",
 "repo": "nvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-BF16",
 "lab": {
  "name": "NVIDIA",
  "country": "United States"
 },
 "released": "2026-10-15",
 "parameters": 74.89,
 "activeParameters": 9,
 "maxContext": 262144,
 "architecture": "KV on 8 of 88 layers (GQA, 2 KV × 128) · 40 Mamba-2 layers · 40 MoE layers (240 to 496 experts) · C-RADIO vision encoder",
 "page": "https://studiotvai.com/vram-requirements/nemotron-3-5-bolt-74b-a9b",
 "calculator": "https://studiotvai.com/llm-vram-calculator?model=nemotron-3.5-bolt-74b-a9b",
 "weightsGiB": {
  "native": 22.43,
  "bf16": 139.49,
  "fp8": 70.75,
  "q4_k_m": 41.85
 },
 "kvCacheGiBPerRequest": [
  {
   "context": 8192,
   "bf16": 0.063,
   "fp8": 0.031
  },
  {
   "context": 32768,
   "bf16": 0.25,
   "fp8": 0.125
  },
  {
   "context": 131072,
   "bf16": 1,
   "fp8": 0.5
  },
  {
   "context": 262144,
   "bf16": 2,
   "fp8": 1
  }
 ],
 "recurrentStateGiBPerRequest": 0.1585,
 "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": 1,
     "bf16": 7,
     "fp8": 4,
     "q4_k_m": 2
    }
   },
   {
    "gpu": "rtx-5090",
    "name": "GeForce RTX 5090",
    "memoryGB": 32,
    "engine": "llama.cpp",
    "byFormat": {
     "native": 1,
     "bf16": 5,
     "fp8": 3,
     "q4_k_m": 2
    }
   },
   {
    "gpu": "rtx-pro-6000",
    "name": "RTX PRO 6000 Blackwell",
    "memoryGB": 96,
    "engine": "llama.cpp",
    "byFormat": {
     "native": 1,
     "bf16": 2,
     "fp8": 1,
     "q4_k_m": 1
    }
   },
   {
    "gpu": "mac-m5-max-128",
    "name": "Mac M5 Max 128 GB",
    "memoryGB": 96,
    "engine": "llama.cpp",
    "byFormat": {
     "native": 1,
     "bf16": 2,
     "fp8": 1,
     "q4_k_m": 1
    }
   },
   {
    "gpu": "strix-halo",
    "name": "Ryzen AI Max+ 395 128 GB",
    "memoryGB": 96,
    "engine": "llama.cpp",
    "byFormat": {
     "native": 1,
     "bf16": 2,
     "fp8": 1,
     "q4_k_m": 1
    }
   },
   {
    "gpu": "dgx-spark",
    "name": "DGX Spark (GB10)",
    "memoryGB": 128,
    "engine": "llama.cpp",
    "byFormat": {
     "native": 1,
     "bf16": 2,
     "fp8": 1,
     "q4_k_m": 1
    }
   },
   {
    "gpu": "h100",
    "name": "NVIDIA H100",
    "memoryGB": 80,
    "engine": "vLLM",
    "byFormat": {
     "native": 1,
     "bf16": 3,
     "fp8": 2,
     "q4_k_m": 1
    }
   },
   {
    "gpu": "h200",
    "name": "NVIDIA H200",
    "memoryGB": 141,
    "engine": "vLLM",
    "byFormat": {
     "native": 1,
     "bf16": 2,
     "fp8": 1,
     "q4_k_m": 1
    }
   },
   {
    "gpu": "b200",
    "name": "NVIDIA B200 (HGX)",
    "memoryGB": 180,
    "engine": "vLLM",
    "byFormat": {
     "native": 1,
     "bf16": 1,
     "fp8": 1,
     "q4_k_m": 1
    }
   }
  ]
 },
 "ggufFiles": null,
 "runsWith": [],
 "verified": null,
 "updated": "2026-10-20T06:01:05.847Z"
}