{
 "id": "beam-1.0",
 "name": "Beam 1.0 501B-A23B",
 "repo": "reflection/Beam-1.0-501B-A23B",
 "lab": {
  "name": "Reflection AI",
  "country": "United States"
 },
 "released": "2026-10-16",
 "parameters": 501.27,
 "activeParameters": 23.4,
 "maxContext": 1048576,
 "architecture": "MoE 256 experts · 9 full-attention layers + 43 sliding-window layers (512 tokens), gated GQA 8 KV × 128",
 "page": "https://studiotvai.com/vram-requirements/beam-1-0",
 "calculator": "https://studiotvai.com/llm-vram-calculator?model=beam-1.0",
 "weightsGiB": {
  "native": 273.47,
  "bf16": 933.69,
  "fp8": 468.29,
  "q4_k_m": 282.01
 },
 "kvCacheGiBPerRequest": [
  {
   "context": 8192,
   "bf16": 0.365,
   "fp8": 0.183
  },
  {
   "context": 32768,
   "bf16": 1.209,
   "fp8": 0.604
  },
  {
   "context": 131072,
   "bf16": 4.584,
   "fp8": 2.292
  },
  {
   "context": 1048576,
   "bf16": 36.084,
   "fp8": 18.042
  }
 ],
 "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": 16,
     "bf16": 48,
     "fp8": 24,
     "q4_k_m": 16
    }
   },
   {
    "gpu": "rtx-5090",
    "name": "GeForce RTX 5090",
    "memoryGB": 32,
    "engine": "llama.cpp",
    "byFormat": {
     "native": 16,
     "bf16": 48,
     "fp8": 24,
     "q4_k_m": 16
    }
   },
   {
    "gpu": "rtx-pro-6000",
    "name": "RTX PRO 6000 Blackwell",
    "memoryGB": 96,
    "engine": "llama.cpp",
    "byFormat": {
     "native": 3,
     "bf16": 16,
     "fp8": 6,
     "q4_k_m": 4
    }
   },
   {
    "gpu": "mac-m5-max-128",
    "name": "Mac M5 Max 128 GB",
    "memoryGB": 96,
    "engine": "llama.cpp",
    "byFormat": {
     "native": 3,
     "bf16": 16,
     "fp8": 6,
     "q4_k_m": 3
    }
   },
   {
    "gpu": "strix-halo",
    "name": "Ryzen AI Max+ 395 128 GB",
    "memoryGB": 96,
    "engine": "llama.cpp",
    "byFormat": {
     "native": 3,
     "bf16": 16,
     "fp8": 6,
     "q4_k_m": 3
    }
   },
   {
    "gpu": "dgx-spark",
    "name": "DGX Spark (GB10)",
    "memoryGB": 128,
    "engine": "llama.cpp",
    "byFormat": {
     "native": 3,
     "bf16": 8,
     "fp8": 4,
     "q4_k_m": 3
    }
   },
   {
    "gpu": "h100",
    "name": "NVIDIA H100",
    "memoryGB": 80,
    "engine": "vLLM",
    "byFormat": {
     "native": 5,
     "bf16": 16,
     "fp8": 8,
     "q4_k_m": 5
    }
   },
   {
    "gpu": "h200",
    "name": "NVIDIA H200",
    "memoryGB": 141,
    "engine": "vLLM",
    "byFormat": {
     "native": 3,
     "bf16": 8,
     "fp8": 4,
     "q4_k_m": 3
    }
   },
   {
    "gpu": "b200",
    "name": "NVIDIA B200 (HGX)",
    "memoryGB": 180,
    "engine": "vLLM",
    "byFormat": {
     "native": 2,
     "bf16": 7,
     "fp8": 4,
     "q4_k_m": 2
    }
   }
  ]
 },
 "ggufFiles": null,
 "runsWith": [],
 "verified": null,
 "updated": "2026-10-17T12:03:02.427Z"
}