{
 "id": "qwen3-next-80b-a3b",
 "name": "Qwen3-Next 80B-A3B",
 "repo": "Qwen/Qwen3-Next-80B-A3B-Instruct",
 "lab": {
  "name": "Alibaba",
  "country": "China"
 },
 "released": "2025-09-09",
 "parameters": 81.32486272,
 "activeParameters": 3,
 "maxContext": 262144,
 "architecture": "48 layers · 12 full attention (2 KV × 256) · 36 linear attention",
 "page": "https://studiotvai.com/vram-requirements/qwen3-next-80b-a3b",
 "calculator": "https://studiotvai.com/llm-vram-calculator?model=qwen3-next-80b-a3b",
 "weightsGiB": {
  "bf16": 148.41,
  "fp8": 74.78,
  "q4_k_m": 44.86
 },
 "kvCacheGiBPerRequest": [
  {
   "context": 8192,
   "bf16": 0.188,
   "fp8": 0.094
  },
  {
   "context": 32768,
   "bf16": 0.75,
   "fp8": 0.375
  },
  {
   "context": 131072,
   "bf16": 3,
   "fp8": 1.5
  },
  {
   "context": 262144,
   "bf16": 6,
   "fp8": 3
  }
 ],
 "recurrentStateGiBPerRequest": 0.072,
 "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": 7,
     "fp8": 4,
     "q4_k_m": 3
    }
   },
   {
    "gpu": "rtx-5090",
    "name": "GeForce RTX 5090",
    "memoryGB": 32,
    "engine": "llama.cpp",
    "byFormat": {
     "bf16": 6,
     "fp8": 3,
     "q4_k_m": 2
    }
   },
   {
    "gpu": "rtx-pro-6000",
    "name": "RTX PRO 6000 Blackwell",
    "memoryGB": 96,
    "engine": "llama.cpp",
    "byFormat": {
     "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": {
     "bf16": 2,
     "fp8": 1,
     "q4_k_m": 1
    }
   },
   {
    "gpu": "strix-halo",
    "name": "Ryzen AI Max+ 395 128 GB",
    "memoryGB": 96,
    "engine": "llama.cpp",
    "byFormat": {
     "bf16": 2,
     "fp8": 1,
     "q4_k_m": 1
    }
   },
   {
    "gpu": "dgx-spark",
    "name": "DGX Spark (GB10)",
    "memoryGB": 128,
    "engine": "llama.cpp",
    "byFormat": {
     "bf16": 2,
     "fp8": 1,
     "q4_k_m": 1
    }
   },
   {
    "gpu": "h100",
    "name": "NVIDIA H100",
    "memoryGB": 80,
    "engine": "vLLM",
    "byFormat": {
     "bf16": 3,
     "fp8": 2,
     "q4_k_m": 1
    }
   },
   {
    "gpu": "h200",
    "name": "NVIDIA H200",
    "memoryGB": 141,
    "engine": "vLLM",
    "byFormat": {
     "bf16": 2,
     "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/Qwen3-Next-80B-A3B-Instruct-GGUF",
  "gib": {
   "BF16": 148.51,
   "Q4_1": 46.62,
   "Q5_K_M": 52.91,
   "Q5_K_S": 51.24,
   "Q6_K": 61.04,
   "Q8_0": 78.99,
   "IQ4_NL": 42.01,
   "IQ4_XS": 39.72,
   "Q2_K": 27.17,
   "Q2_K_L": 27.24,
   "Q3_K_M": 35.67,
   "Q3_K_S": 32.21,
   "Q4_0": 42.2,
   "Q4_K_M": 45.17,
   "Q4_K_S": 42.38,
   "UD-IQ1_M": 22.73,
   "UD-IQ1_S": 21.33,
   "UD-IQ2_XXS": 24.41,
   "UD-IQ3_XXS": 30.82,
   "UD-Q2_K_XL": 28.06,
   "UD-Q3_K_XL": 33.19,
   "UD-Q4_K_XL": 42.9,
   "UD-TQ1_0": 19.06,
   "UD-Q5_K_XL": 52.77,
   "UD-Q6_K_XL": 63.81,
   "UD-Q8_K_XL": 86.68
  }
 },
 "runsWith": [
  "vLLM",
  "SGLang",
  "llama.cpp (Ollama, LM Studio)",
  "MLX (Mac)"
 ],
 "verified": null,
 "updated": "2026-11-06T21:02:26.190Z"
}